By Paul Homewood
We have already seen that the GISS temperature record in East and North Africa is utterly worthless, and probably grossly misleading, being based on a tiny number of stations, all of which are urban or airport based.
But if there is one place on the continent where we might expect a decent selection of high quality sites, it would be South Africa. However, it appears things are little better there.
According to GISS, the country has had about 0.5C or so of warming, compared to 1951-80.
There are twelve sites currently reporting on the GISS database, but only three which match the GISS criteria as “rural”, i.e. with a Brightness Index of 10 or under. As we see below, two of these are busy airport sites, linked to large towns.
1) George Airport
The station is based at this airport:
According to Wiki, the airport was built in 1977 (coinciding with the resumption of temperature records, after a gap), and has increased its annual passenger numbers from 154,000 in 2003 to 560,000 last year.
Furthermore, the town of George, only 3 miles away, is a large, developed one, with a population of 157,000.
To make matters worse, the early, pre-airport part of the temperature record cannot be compared with the most recent segment because of the substantial change in siting, and the huge gap in the record between 1963 and 1986.
Despite obvious UHI bias in recent years, the temperature record shows no warming since 1986.
1) The current site is not a suitable one for climate purposes.
2) The long term record is unreliable because of the large gap, clear step change in temperatures and location changes.
Again, Upington is airport based. Although not as busy as George, the airport serves as an international cargo hub, with over 3000 flights a year.
The airport is only three miles from the town of Upington, which has a population of 75,000.
The airport was opened in 1968, and is being into a major cargo hub.
1) Again, as an airport it is unreliable for climate purposes and records prior to that are not consistent.
Calvinia seems to be a more reliable site, as it is not airport based and has a smaller population of 10,000.
As is the case with Upington, there is a sharp decline in temperatures in the 1960’s and 70’s, with signs of temperatures prior to that being similar to recent years. Unfortunately, Calvinia only has four years of temperature records between 1943 and 1959. The lack of this data could be giving a false impression of trends.
It’s not exactly scientific, I know, but the average of the four available years in the 1940’s, (1945-7 and 1949), average 16.34C. The average from 2001 to 2013 is 16.38C
1) The lack of more than a handful of records before 1959, together with a sharp decline in temperatures shortly after, means that the trends are unreliable.
Out of the other nine stations:
1) Six are airport based
2) One is at Pretoria, with a population of 741,000.
3) Another is at Bloemfontein, with 256,000 inhabitants.
4) And the other is at Dev Aar, with 30,000.
Interestingly, a batch of genuinely rural stations, many at agricultural stations, stopped recording in 2002, most with records back to 1960. Apart from the 1970’s dip, there seems to be little long term trends up to 2001, as the examples below show.
The bottom line is that, even if our three “rural” sites are giving unbiased trends, to base the climatic trends of a country as big as South Africa, and likely well beyond the borders, on just three sites surely cannot be depended on to give robust results.
As I have said before, we really have little idea of what temperature trends have been in Africa. To guess and pretend that we do is not science.