Rothamsted Addendum – A Study into rising mean minimum temperatures using comparative data.

51.80663 -0.36017 Met Office CIMO Assessed Class 1 Installed 1/1/1872

I originally reviewed Rothamsted in the early stages of my resumption of the Surface Stations Project. This is not a further study of the site which I feel, despite some others’ views, is a very good quality site. The site’s custodians (Rothamsted Research) operate with the highest scientific integrity and I have not discovered any adverse issues with the quality of the data that they produce. I have, however, recently expressed concerns with the manner in which the Met Office analyses and presents data from all its sites. The Met Office presentation has resulted in an ongoing misrepresentation of night time minima detailed here.

Either by happenstance, or possibly the very high diligence of Rothamsted Research staff, their site offers comparative data to test my assertion that the Met office has/is presenting data known to support a warming temperature trend where it may not actually exist.

Several people have requested clarification of my assertion that daily mean minimums (particularly winter periods) can be distorted by readings frequency thus I will recap some basic points for clarity. Manual climate reporting weather stations take observations at 09:00 GMT every day of the year. Included in these readings are daily maximum and minimum air temperatures. These two readings are then added together, divided by 2 and thus derives the daily mean temperature. I fully explain this unusual and rather crude meteorological averaging system here.

To visually explain the procedure further, below is an image that most will be familiar with. This is a basic “Six” thermometer originally designed in 1780 by the British scientist James Six. {n.b. Met office instruments are obviously much more sophisticated but this one is adequate to illustrate the point}

Maximum and minimum readings are noted from the bottom points of the two small iron rods (“Riders”) which are either pushed by, or retracted from, the thermal expansion/contraction of the internal liquid medium in response to temperature changes. The observer, having noted the readings at 09:00 GMT will then “reset” the thermometer by using a magnet to move the riders back down to sitting atop the liquid column. A small side filament attached to each of the riders stops them from naturally falling under gravity, the magnet overcomes this resistance. Now reset, the thermometer is left alone until 09:00 GMT the following day with the process repeated every day thereafter. Only one maximum and one minimum reading are archived.

As I highlighted in my earlier report, the usually coldest time of the day is approximately half an hour after sunrise.

The most southerly point of mainland Great Britain is Lizard Point at 49° 57′ 30″ N. – for comparison this is further north than anywhere in the 48 contiguous states of the US. The most northerly UK weather station is Balta Sound at 60°44’56.1″N.

The point I am emphasising is that the UK is a high latitude country which results in it being a relatively “dark” country in winter. – {Ed note: A simple point that seems to be lost on advocates of solar photo-voltaic electricity generation}. The long term Morpeth Cockle Park weather station at 55°12’44.8″N actually represents the mid north south latitude point of the UK weather station sites excluding the Channel Isles. At Morpeth sunrise is later than 08:00 GMT from mid November through to Early February with the latest point at 08:32 GMT. This results in minimum daytime temperature recording being frequently taken at, or close to, the coldest time of day. All UK manual reporting stations are subject to this over-emphasis on a “cold” reporting time, particularly in winter.

Referring back to the observation process, the reset occurs close to the identified coldest time of day and thus, even if the temperature rises throughout the following 24 hours, that reset minimum can be recorded the following day as well – a single cold night is often recorded twice i.e. over two days.

If readings are taken more frequently and at later times of day, the reset process will occur later in the day. The type of thermometer used (LIGT or PRT) is not relevant as any further readings being noted will reset the readings. Typically on automation (but also at any more frequently human observed stations) the additional archived readings are taken/noted for 21:00 GMT being after the generally warmer daytime hours and before the nighttime drop. In this situation the double recording of one night’s colder temperature noted at manual/once daily reporting sites does NOT happen.

The issue of how Rothamsted enters in to this debate is demonstrated by its change over from manual to automatic readings. When the Met Office automates a site there is a changeover from LIGT to PRT readings with the latter unit’s observations being noted twice daily. Normally the changeover involves a complete immediate swap out of equipment, and often screen, though occasionally two separate systems run concurrently to cross reference readings for accuracy. Rothamsted did NOT convert over like that. In fact it ran two independent systems that had data recorded for just over 4 years and even thereafter Rothamsted Research themselves continued the manual reporting for another 3 years. The Met Office did not pass these later readings to the CEDA archives but Rothamsted did themselves retain them.

The headline image shows two independent Stevenson screens – the original manual and the newer automatic which existed between 1999 and 2007. Up to 28th November 1999 only once daily 09:00 manual readings from the original screen are shown below.

Column A indicates day and reading time, column D indicates “24” i.e. once daily reading, column F indicates the reading type i.e. manual on reporting form DLY3208. Columns I & J are maximum and minimum air temperatures respectively.

From 1/1/2004 only automatic twice daily readings are shown ( Note the last manual DLY3208 maximum reading on the 1st – manual maximum readings are attributed to the day before i.e. 31/12/2003 so this was supplied for annual completion of 2003)

Now we have 12 hourly readings solely under AWSDLY (Automatic Weather Station Daily).

The interim period, however, looks like this, for which I have shown both the first 10 day period in 1999 below and the last corresponding same period for 2003 under that.

1999

And 2003

The first thing to notice are three daily readings, two shown as AWSDLY at 09:00 and 21:00 and one DLY3208 at 09:00. This indicates that there are two minimum readings both taken at the same time (09:00 GMT) from equally well calibrated instruments. There should be NO VARIANCE between these two readings if they were recording the same thing but as stated above they are not always. The DLY3208 is showing the minimum since reset 24 hours earlier (close to the coldest time of day) whilst the AWSDLY is recording the minimum since 21:00 GMT – only 12 hours previous.

So is there much variation or not? It is my intention to sample the entire 4 year period shortly with particular reference to any seasonal variation and relation to sunrise timing. For now I have just quickly scanned through the above 10 day start and end periods with alarming results.

Minima 1999 (degrees celsius)

Date………………………………DLY3208………………………….AWSDLY………………Variation AWS above DLY

30/11/1999……………………………..-0.5…………………………………-0.2…………………….+0.3

1/12/1999……………………………….2.1………………………………….2.1………………………Same

2/12/1999……………………………….4.4…………………………………..4.4……………………..Same

3/12/1999……………………………….2.0………………………………….6.0………………………+4.0

4/12/1999………………………………1.3………………………………….3.0………………………+1.7

5/12/1999………………………………0.5………………………………….0.4………………………-0.1

6/12/1999……………………………..0.9…………………………………..4.2………………………+3.3

7/12/1999……………………………..7.8…………………………………..8.1……………………..+0.3

8/12/1999……………………………..2.7…………………………………..2.6……………………..-0.1

9/12/1999……………………………..4.1………………………………….6.3……………………..+2.2

Period mean……………………….2.53……………………………3.69…………………..+1.16

And for the same period 4 years later.

30/11/2003…………………………4.6………………………………..4.5……………………….-0.1

1/12/2003…………………………..5.5………………………………..7.2………………………+1.7

2/12/2003…………………………..5.4………………………………..5.3………………………-0.1

3/12/2003…………………………..6.0……………………………….6.2………………………+0.2

4/12/2003…………………………..6.5……………………………….7.4……………………..+0.9

5/12/2003…………………………..6.7……………………………… 6.9……………………..+0.2

6/12/2003…………………………..1.1……………………………….0.9………………………-0.2

7/12/2003…………………………..0.8……………………………..0.6……………………….-0.2

8/12/2003………………………….-2.8……………………………..-2.5……………………..+0.3

9/12/2003…………………………..1.9………………………………4.6……………………..+2.7

Period mean……………………..3.57…………………………4.11………………….+0.54

My apologies for the crude presentation and small period initial sampling. The entire period will be fully analysed with significantly better presentation when available. However, it is immediately evident that the Rothamsted data clearly demonstrates the effect that changing the reading frequency has removed the often cold overnight near minimum double counting effect. This in turn has had a significant effect on raising the overall average mean minimum and will thus increase the overall mean daily temperature…..WITH NO ACTUAL WARMING WHATSOEVER.

Dr Eric Huxter is thoroughly analysing data from Rothamsted with reference to PRT recorded temperature spikes relative to CIMO ratings. He recently graphed temperatures recorded there noting an emphasis on the LIGT reporting period with the subsequent PRT reporting. in addition to the effects he is noting can be added this non meteorological effect solely caused by changes of observervation protocols.

From the above, the noted mean daily temperature appears to show a step change coincident with both instrument change (LIGT to PRT) and also consequent observation recording protocol. Whether or not the two changes are complementary, opposed or any other possible combination is not yet known but clearly some effects are taking place which warrant deeper consideration. As Eric points out though it seems the Met office conveniently overlook any potential flaws that supply data to prove their preset “cause”.

Source: https://tallbloke.wordpress.com/2025/09/26/rothamsted-addendum-a-study-into-rising-mean-minimum-temperatures-using-comparative-data/