What was observed
The minimum thermometer reaches a record
The station register gives −89.2°C at 02:45 UTC on 21 July. Turner and colleagues’ later reconstruction places this minimum at the end of roughly ten days of substantial cooling, rather than treating it as an abrupt isolated plunge.
Basis: World Meteorological Organization [1] · Turner and colleagues; British Antarctic Survey and Arctic and Antarctic Research Institute [2] · British Antarctic Survey [3]
Selected observations & analyses
These entries preserve the source’s reported quantities or analyses. They are not a newly processed observational dataset.
| Evidence | Reported value or finding | Provenance |
|---|---|---|
| Extreme minimum | −89.2°C on 21 July 1983 | WMO extremes archive, supported by the station register examined in Turner and colleagues’ 2009 study. |
| Scheduled observation | −88.8°C at 00:00 UTC on 21 July | Turner and colleagues’ analysis of the six-hourly station observations; the scheduled sample preceded the lower intervening minimum. |
| Preceding weather reports | No cloud or diamond dust reported over the station for about a week before the minimum | Turner and colleagues’ reconstruction using the original meteorological register; a reported condition relevant to radiative cooling. |
The atmospheric mechanism
During polar night there is no local solar input to balance the surface’s infrared energy loss. Clouds and atmospheric water vapour can return infrared energy toward the surface, so a sustained clear, dry atmosphere permits stronger cooling. The snow surface then cools the air immediately above it, producing an inversion in which temperature rises with height through the lowest layer.
Isolation matters as much as darkness. The research reconstruction found little replacement by warmer maritime air and relatively weak surface winds for Vostok. Weak mixing limits the transfer of warmer air downward, allowing the shallow cold layer to persist. The high plateau and cold atmosphere provide the background setting, but the record required a particular sequence of weather rather than elevation alone.
An observation schedule also shapes what enters a dataset. A reading taken every six hours can miss the coldest moment between reports, while a minimum thermometer preserves that lower value. Neither number needs to be wrong. A careful comparison keeps the sampling interval beside the measurement and avoids replacing a daily extreme with the nearest convenient timestamp.
Basis: World Meteorological Organization [1] · Turner and colleagues; British Antarctic Survey and Arctic and Antarctic Research Institute [2] · British Antarctic Survey [3]
What the evidence does not establish
This is a near-surface air-temperature record, not a satellite estimate of snow-surface temperature. The 2009 analysis combines observations with reanalysis and physical interpretation; its process diagnosis is not another thermometer measurement. The selected date marks the record minimum, not the beginning of the entire cooling period. The location follows BAS coordinates because source summaries differ in their station metadata.
Ask what supplies or removes heat over several days, and keep the timing of a scheduled observation separate from the interval’s recorded minimum.
Sources & provenance
- Global and Antarctic lowest-temperature record World Meteorological Organization · Record evaluated in 2011; archive entry undatedReference only; original paraphrase; no source media reproduced.
- Record low surface air temperature at Vostok station, Antarctica Turner and colleagues; British Antarctic Survey and Arctic and Antarctic Research Institute · 2009-12-16Reference only; original paraphrase; no source media reproduced.
- Antarctic Manned Meteorological Stations British Antarctic Survey · Publication date not stated; retrospective event accountReference only; original paraphrase; no source media reproduced.
Explanations and diagrams are original educational interpretations. Source institutions have not endorsed or reviewed this publication. How we distinguish evidence from models.