« The Potency of Cheap Music | Main | First they came for the smokers… »

Rothamsted Temperature Record Problem?

The Rothamsted Research Centre is one of three temperature recording centres of record for the official Central England Temperature. I can't spot the Stevenson screen on Google Maps but the research station is still fairly large and green. So is it a reliable source of data?
The driver of the increasing warmth in the UK is mainly the increase in the T min (Minimum Temperature - we aren't getting the cold to bring the averages down).
Here is the T min series:


Wow, quite a jump in the Tmin for Rothamsted (the solid red line).
As the Met Office notes - "the coldness of Ringway around 1993 the warmth of Rothamsted around 1993 and 2001, while disconcerting, may not be outliers in view of the expected uncertainties. That Rothamsted T min is not fully supported by Cambridge T min suggests calibration or siting biases rather than real spatial variations of min temperature anomalies. If any biases can be ascertained following further investigation of Ringway and Rothamsted, then the CET series should be adjusted accordingly." Source

I can't find any evidence that they have done any further investigations or adjustments, so the extra half a degree(?) from Rothamsted still feeds into the data.

Just eyeballing that, does it look right to you? I have no idea what is going on, is it correct or whether there is a software problem such as Steve McIntyre found with the US records, or a siting bias as the Met Office suggests or maybe something else.

Just north and north west of Rothamsted lies the Bedfordshire brick fields, an average distance would be about twenty miles away. There used to be over 130 chimneys there belching out sulphurous fumes. As we know this sort of pollution cools the atmosphere and is bad, so they have all been knocked down apart from one or two, which try to be as non polluting as possible.
This is the area in 1974:

Could this be the source of the anomalous jump in Rothamsted's temperature average, which fuels the jump in the UK record?


Always an alternative explanation.

Mr England Project's Mazda?

(He lives in Harpenden)

Looking at the Rothamsted websites, there has certainly been building work, but I couldn't spot where the louvred box is. If it is now surrounded by conference centre...

I worked/studied at Rothamsted for 4 years (1985-1989) and there was significant development of the site, in terms of new buildings, which has continued since then. There was also significant development of the surrounding area as Luton (7-8 miles to the north) experienced a housing boom right up to the county border, although Harpenden itself didn't experience very much house building during this time. I don't know where the weather station was/is situated and there is a lot of experimental agricultural land so a well-sited station would be reliable from the point of view of siting effects, but there could still have been a urban effect from the expansion of Luton. How far would an urban heating effect reach?

Hey, let's put the thermometer down on the concrete sidewalk in front of the kewl new giant reflecting parabola building...

My name is Willis Eschenbach, and I am the author of the graph shown on your "Central England Temperature - A Waste Of Time" thread at http://www.anenglishmanscastle.com/archives/004431.html

Thank you for providing my work to a larger audience.

I am interested in a further investigation of the temperature records shown in your graph above. What is the source of the raw Tmin data for the graph?

With appreciation,


The graph came originally from Parker, D.E. and Horton, E.B., 2005: Uncertainties in the Central England Temperature series 1878-2003 and some improvements to the maximum and minimum series. International Journal of Climatology , 25, 1173-1178. The curves show DIFFERENCES from CETmin so when one of Rothamsted, Malvern, Ringway and Squires Gate goes up the average of the rest has to go down. So the variations are accentuated. The slight overall rising trend in the curves since the 1980s is because no urban warming adjustments have been applied to these station data whereas the CETmin has been corrected by up to 0.4 deg C. The urban warming adjustments to CETmin mainly related to Malvern and Ringway data, not Rothamsted.

Developments in Bedfordshire and Luton are too far away to have a significant effect on Rothamsted. But we have tested the hypothesis of local urban warming as follows. Urban warming is most noticeable on calm clear nights and much less evident when it is windy or cloudy or both. So (in the absence of reliable cloud-amount data) we compared trends of Tmin at Rothamsted on windy and calm nights. The calm nights showed no relative warming, which they would have done if there had been a local urban warming influence.

David, thank you kindly for the response. Tim had been kind enough to point
me to your interesting paper, from which I extracted the minimum temperature
data shown in the graph above.

What I found curious about the temperature data was the correlation between
the five stations shown above. Of the ten cross-correlations, three are
actually negative, which was a great shock. In addition, half are below
0.20, and the highest is only 0.57.

Hansen has commented extensively on the correlation between widespread
temperature stations. For example, Fairbanks and Anchorage (Alaska) are 420
km. apart, but their correlation is 0.87.

I'd be interested in your thoughts about why the station cross-correlations
are so bad. For me, it's a very curious thing, worthy of explanation ... and
until I figured out the answer, I'd sure put wide error bands on any average
using those stations.

The correlations in the diagram are low because the plots are of DIFFERENCES of station anomalies from the anomalies of Central England temperature. The correlations of actual monthly mean station anomalies range from 0.89 in May (when atmospheric circulation has the weakest unifying influence) to 0.98 in January. For monthly average Tmin they are a little lower but still high, ranging from 0.87 in August to 0.97 in January. The differencing removes the unifying influence of large-scale atmospheric circulation.

David Parker, thank you for your prompt reply. I would be most interested in your further comments on my observations in the thread at http://www.anenglishmanscastle.com/archives/004482.html

All the best,



There's the temperature station - it's 8 meters from that metal tower. You have to look south of the research centre to find it.

Here's the new Hadley CET station at Pershore

I can't locate Stonyhurst - perhaps because both the bingmaps and google images are not recent enough.

Post a comment