RUTI: Western Europe Rural Temperature Trend
(Update 6/7 2011, see end of article)
- Searching for the long rural stations in Europe. Are almost all rural temperature data limited so much, that we in reality are building temperature data mostly on urban data?
To what extend are temperature indexes from Hadcrut, Giss, GHCN etc based on rural data?
After making the writing articles like “RUTI South Africa” and “RUTI Turkey” etc. where all rural datasets have been ut down to just the period 1960-90, I got the impression that the share of long rural stations is perhaps more limited than most believe?
Lets check it out for Europe.
A “long temperature series” would be a station from where we have most years 1900 till 2010 included.
Fig1. In fact, we can loosen this definition somewhat, because the interesting part is to examine the warm peak 1925-1940 with the warm peak of the most recent years 1995-2010:
Fig2. We can loosen our demands further: We only need to see 10 of the 15 years for both periods 1925-40 and 1995-2010, since this will often give a fair chance to compare the 2 warm peaks:
Fig3. In general I am looking for stations representing the bulk of the European land mass, not extreme locations like icy mountain tops (see RUTI Alps) or smaller Islands (to a high degree representing Marine Air Temperature trends).
We can expand our definition further: If a temperature station is near a city but clearly outside a city (perhaps 1 mile from urban area) we can accept the station too.
So lets see what long rural temperature series we have available from GHCN:
Fig4. Blue starts: Temperature stations fulfilling criteria’s for “long rural temperature series” as defined above. Green stars: Stations almost fulfilling demands except we have more than 10 years in the period 1925-1949 in stead of the shorter period 1925-40. Red stars: Stations fulfilling demands except location is on mountain top (see RUTI Alps).
Great Britain + Ireland : Plenty stations meet the above criteria’s for long rural temperature series, see RUTI Great Britain.
Orange areas on map above: Countries without long rural temperature series meeting criteria’s, at least in area shown.
The number of rural stations generally in GHCN is irrelevant. For example in Turkey there are around 100 rural stations but non of these are used outside the interval 1960-90 (See RUTI Turkey). So although almost half the Turkish GHCN stations are rural, not a single Turkish rural temperature station has any effect on the long trends, the trends that matters.
So to answer my original question:
"To what extend are temperature indexes from Hadcrut, Giss, GHCN etc based on rural data?"
I would have to say: Very little.
This was the main point of this article, but lets move on and check out the European long rural temperature stations of Europe as pointed out above:
Fig5: 4 Western European long rural temperature series. I don’t have to mention the obvious similarity between these data series when no urban influences, marine stations or mountain peak stations are used.
Could the similarity be a coincidence? I wont answer that question since my answer is no better than any readers answer. I think what all could agree is: GHCN Should release all data, urban and rural series for France, Germany, Poland, Italy… all the world. If this does not happen, people might suspect that trends like the above are what GHCN is holding back from the public.
Fig6. Finally NASA´s GISS uses 1200 km radius around their stations when mapping temperature trends. I tried to apply this on the 4 long rural temperature stations in the above area of Western Europe. If NASA/GISS are correct, and it is recommendable to use 1200 km radius for temperature stations, then it might not be such a coincidence that the temperature trends showed some similarity.
But obviously, to conclude much more we need all temperature data released.
For Italy, I found this near rural couple, both from Catania:
And check out their combined temperature series:
And then lets add the combined (averaged) Catania to the other rural western Europe data series: