Nonsense, there's no "scientifically universally accepted definition" of the 20th Century Average.
OK then, find me a single use of "20th century average" which includes data from before 1900 or after 2000. It doesn't have to be to do with temperature, it doesn't even have to be from a scientific source. Just one example of anyone anywhere using it the way you wrongly thought NOAA was using it.
The 19th Century average, which is in the paper by the way is the years 1880-1900, not 1800-1899 like you keep insisting.
I never insisted anything of the sort.
It's the average of all the data that's available for the 19th century, as has been explained to you multiple times. It just so happens that, for global temperatures, that data begins in 1880.
For any quantity where data is available for the whole of the 19th century, the 19th century average would mean the average from 1801-1900.
The 21st Century average, which is also referenced in the Reynolds paper is the years 2000-2006.
It's the average of all the data that's yet available for the 21st century, as has already been explained to you multiple times. Prior to 2000, anyone who referred to the 20th century average would likewise have meant the average for the century
so far.
Clearly there's not "universally scientific definition" of the "xth Century Average".
Yes there is, it's the average from (x-1)01-x00 or, where the available data covers a period shorter than that, the average for the years for which data is available. It never, ever, covers a period of more than a century.
I thought the switch to version 3 meant the so called "full period" was being used and NOAA simply didn't update their website to reflect this.
You thought very very wrong. Again: improvements to or extensions to the NOAA data set is never ever going to change the way they define "20th century average". Why should it? It's just a convenient, universally understood (by everybody but you) baseline against which to calculate temperature anomalies. For some things NOAA chooses a different base period, for example on the temperature anomaly maps they use a base period of 1961-1990. They're not going to change that one every time they improve or extend their data set either.
And if they did, there's no real reason other than pendantry to do so. The bulk of the average is the 20th Century, the few years on either side really don't mean squat
As it happens, this particular dataset only covers a few years either side of the 20th century, but that doesn't mean that it would ever have been correct to use "20th century average" to describe the average over the whole period. If NOAA had ever done so they would have been a laughing stock.
And why would temperature anomalies be calculated wrt the average over the whole period for which data is available anyway? Every time another full year's data became available all the previously published temperature anomalies would have had to be recalculated, which would be absurd.
It makes sense to pick a fixed period in the past for which data is available and calculate anomalies wrt the average of it, and continue to use the same base period as more data becomes available, and improvements are made to the analysis of existing data.
.
!
!
).