WunderBlog Archive » Weather Extremes

Category 6 has moved! See the latest from Dr. Jeff Masters and Bob Henson here.

New Cold Record for South Africa?

By: Christopher C. Burt, 10:07 PM GMT on August 25, 2013

New Cold Record for South Africa?

The South African Weather Service reported a low temperature of -19.5°C (-3.1°F) at Buffelsfontein on Friday August 23rd. If verified this would be the coldest temperature ever measured in the country.



The preliminary national weather extremes table for South Africa on Thursday-Friday August 22-23. Table from the South African Weather Service.

The previous record cold for South Africa was -18.6°C (-1.5°F) set on June 28, 1996 also at Buffelsfontein. This site is apparently a large farm of some sort near the town of Molteno (not to be confused with the Buffelsfontein Game and Nature Reserve near Cape Town). It has an elevation of 1780 m (5875 feet). (NOTE: yesterday I posted the wrong location of the site).



Location of Molteno which is the town nearest the farm Buffelsfontein in South Africa's East Cape Province. Map from Wikipedia.

However, this reading is most likely an error. Although Buffelsfontein is known as having measured the coldest temperature in South Africa previously, this occurred during an exceptional cold spell whereas on Friday temperatures at most other locations in South Africa were not abnormally cold and, according to temperature climatologist and temperature detective Maximiliano Herrera, no other main site in the country recorded a temperature lower than -8.0°C (17.6°F). Max says he has observed some irregular temperatures coming from a number of sites in South Africa and Namibia recently with anomalous temperature departures of as much as 10°-15°C (18°-27°F) below reality. This may be the case with Buffelsfontein, especially since no other site in the country has come within 10-15°C of their record low.

Also, there has been no mention of this ‘new national record’ in the South African media, which is odd. So, for the time being, I don’t think we can accept this as an official record pending further verification from the South African meteorological authorities.

Christopher C. Burt
Weather Historian

Extreme Weather Cold

The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.

Reader Comments

South Africa has mountains over 11,000 feet high. If they established an observation station above 3,000 meters I expect they would get a new national record low easily. Lesotho too.
Wouldn't necessarily assume that - for extreme low minimum temperatures you need a location favourable to cold-air drainage and development of inversions, i.e. a valley site (preferably one high enough to be above the snowline, at least occasionally). I doubt a 3000-3500 metre high mountaintop at South African latitudes would produce minima below -19.5.

Another thing which makes me sceptical of this report is its timing, very late in winter. As a subtropical country South Africa warms up quickly in spring, and August (especially late August) is significantly warmer on average than June and July (in fact at some sites it is almost as warm as May).
That is correct Blair. I just returned from Jackson Hole, Wyoming on vacation. The valley floor at 6000'-7000'has recorded much colder temperatures than Rendevous Peak (at 11,200'). Moran at 7000' registered -63F on Feb. 9, 1933 and Jackson at 6000' -52 in 1924.

Rendevous Peak has only been open since the mid 1970s but no temperature under -37F has been recorded there since 1975 but the winter minimums in the valley are always lower than the peak itself.
Good point by both of you. I hadn't realized the importance of cold air drainage.
Quoting 2. blairtrewin:
Wouldn't necessarily assume that - for extreme low minimum temperatures you need a location favourable to cold-air drainage and development of inversions, i.e. a valley site (preferably one high enough to be above the snowline, at least occasionally). I doubt a 3000-3500 metre high mountaintop at South African latitudes would produce minima below -19.5.

Another thing which makes me sceptical of this report is its timing, very late in winter. As a subtropical country South Africa warms up quickly in spring, and August (especially late August) is significantly warmer on average than June and July (in fact at some sites it is almost as warm as May).


This is the case. But in South Africa southern highlands cold winter nights with still and dry air are NORMAL conditions in winter, not an exception.
Every single winter we have dozens of days/nights with these typical anticiclonic conditions with very strong irradiation and huge thermal daily excursions.
So we would expect to set a record with a very low isotherm PLUS these typical and necessary conditions to have strong night heat loss by irradiation.
The second condition alone would just cause one of the many freezing nights perhaps between -5C and -10C typical of that cold spot.
If there is absolutely nothing exceptional these days, how can the final result be exceptional ?
Everything should have a phisical explanation, a cause.
If Buffelsfontein can record almost -20C in nights like these ones, so in 20-30 years time it would surely had the chance to approach -30C.
Unless -this is another possibilty- the station was lately moved to a nearby doline. Well, in this case perhaps the reading could be possible, but it wouldn't be very representative as "all-time national record". In this case we can expect new records coming soon in the next winters.
I think the Buffelsfontein you're looking for is neither of those; it's a farm located near the town of Molteno at about 31°S, 26°E. See http://www.news24.com/SouthAfrica/News/Which-is-SA s-coldest-town-20100720 which mentions a farm named "Buffelsfontein".
Quoting 6. Soap2:
I think the Buffelsfontein you're looking for is neither of those; it's a farm located near the town of Molteno at about 31°S, 26°E. See http://www.news24.com/SouthAfrica/News/Which-is-SA s-coldest-town-20100720 which mentions a farm named "Buffelsfontein".


Yes this is the correct location. It's located at 1780m aslm, the map is wrong.
Anyway, the reading is very unreliable.
There wasn't any exceptional or even near rare event, nothing. Absolutely normal winter conditions, in fact in most part of the country with above average temperatures.
Quoting 1. BaltimoreBrian:
South Africa has mountains over 11,000 feet high. If they established an observation station above 3,000 meters I expect they would get a new national record low easily. Lesotho too.


There is one high elevated station in lesotho which recorded -21C in 1967.
But I am not sure about its exact location, I assume it wasn't exactly on a top of a mountain, instead in a high elevated valley.
1967 and 1968 saw two consecutive EXTREMELY cold events,with isotherms as low as -5C in Southern South Africa and the 0C up north to the Tropic Line.
These days we had a +5/+6 in the Southern parts of South Africa: a joke. Mild conditions. Very dry and still air, yes and freezing nights (0C to -7C) and warm days (16C to 25C) in the highlands above 1000m . If that place records -20C with these conditions, in 1967 or 1968 (almost -13C in Zimbabwe !) it would make -30C easily.... Impossible unless it's a sinkhole.
Thanks Christopher!
Hi everybody,

I agree that this reading is spurious. While the GFS model analysis of 2-meter temperatures is not a high-resolution display, it nonetheless indicates the general regional pattern. I include the 06Z GFS model analysis of 2-meter temperatures below (isotherms drawn every five degrees Celsius; 06Z is 8 A.M. local time; larger image). While the impact of microclimates is not reflected in this analysis, I find it very difficult to believe that the nighttime temperature got down to almost minus 20 Celsius. Really, it shouldn't even be close.



I point out that the GFS model analysis of mean sea-level pressure at 06Z (below; larger image) suggests a night where microclimates would likely be colder than the 06Z temperatures indicate (high pressure fosters light winds and a clear sky), but it's a big stretch for me to believe this reading is accurate.

weatherhistorian has created a new entry.
The South African Weather service posted this recently:
http://www.weathersa.co.za/web/images/PDF_docs/ne w_minimum_temperature_record_sa.pdf
I notice that Queenstown has reported -19.3 today (from ogimet, haven't confirmed with SAWS). This appears to be a new station established late last year and has had many extreme low temperatures already this year. I'll follow up with my South African contacts.

(Update: the SAWS website says it was -19.3 today - at Buffelsfontein. This suggests that Ogimet has an error in its station directory and the "Queenstown" data are actually from Buffelsfontein).