Keele DCNN4167 – The issue of twice daily readings being ignored and a startling revelation.

52.99892 -2.27023 Met Office CIMO Assessed Class 4 Installed 1/1/1951

Keele weather station lies within the extensive grounds of the University just 2 miles to the south west of central Stoke on Trent. Although from aerial imagery the site does not look particularly compromised and appears to be in relatively open country, when viewed at street level the size and nature of the trees becomes all to apparent. The Class 4 status and attendant lack of representation of a wider area is warranted. However that is the secondary issue here, yet again data adjustment rears its ugly head.

Springpool Wood is an extensive area of forestry that extensively shelters this site from winds from the north east right around clockwise to the south. University buildings provide additional shelter from the west around to the north east. These issues do not improve the site but perhaps more important is the sheer size of the nearby specimen trees. The main one in the above image is a copper beech and is likely in the region of 40 metres (130 feet) tall. Such trees can cast enormous and deep shadow rendering readings inaccurate.

The site has been operational since 1951 but archived readings only start from 1959. The site was originally manually observed as all others were at that time then converted to automatic readings from 2007.

This early years gap in transcription at many stations has often puzzled me so I decided to investigate the issue of these omissions with very enlightening results. Below is an extract from the first transcribed years with Fahrenheit readings crudely converted to Celsius as I noted in my Wye report on converted data accuracy.

Note how consistently cold (column J) this period was in being an inland site. Several consecutive days readings are exactly the same. Keele continued to record air frosts up to mid May that year. What is more relevant to note is that the reading frequency is shown as just once daily at 09:00. For whatever reason the current reading protocol since automation is at 10 minutes before the hour at 08:50 and 20:50.

2010-09-28

Current

OBSERVING PRACTICE

HOURLY READINGS ARE AT HH-10 RATHER THAN HH+0. DAILY READINGS ARE AT 0850 AND 2050 UTC RATHER THAN 0900 AND 2100 UTC

This really is no particular issue with ten minutes differential not causing any significant difference to a PRT/Data Logger recording system taking readings every minute of the day. As I have mentioned in multiple recent posts, however, the transition from once daily to automation does cause significant problems. In fact it is not just automation that causes the appearance of warming by removing the previous defect of just once daily manual readings – any increase in reading frequency even by LIGT eliminates the problem as well. In manual observations the “riders” are reset on recording and thus avoids carry over of previous mornings readings as miniums even if the temperature warms up. So what was the reading protocol at Keele for those non transcribed years of 1951 to 1958?

The site was originally established taking TWO readings per day at 09:00 and 15:00. Checking all the manual records showed the same process – here is 1958.

At this point pure logic must take over. Readings were established as twice daily for 8 years from installation proven by sight of the manual records. Data from 1959 onwards is only shown as once daily taken at 09:00. Taking the later in the day reading resolves the carry over problem of cold weather events and this site is clearly regularly prone to such events. What exactly does this mean? Did the observations protocol change dropping the 15:00 reading or were twice daily readings continued but resolved down to the highest and lowest of the two being recorded?

The answer of which I am 99% certain (but cannot yet definitively prove) is that the Met Office opted to standardise manual observations readings to just once daily along with other latter 1950s/60s issues such as the conversion to Celsius. This is a fundamentally crucial change in that over recording of cold events was introduced in some sites where it previously had not occurred.

Now contrast this with my other example site quoted at Wye College. There the readings were always taken once daily at 09:00.

This continuity of observations protocol allowed direct comparison over time with Dave Woolcock producing graphs like this indicating very little warming. This is average daily minimum.

So now to my personal startling revelation. In the responses to the Wye post where I revealed the Talkshop’s first attempts at a serious historical reanalysis, I was adversely responded to claiming I was doing it all wrong. Apparently straight comparison of temperature change over time was incorrect and Dave and I should have followed some rather esoteric route of comparing all our data to what seemed a somewhat arbitrarily chosen “reference” period of 1961 to 1990. The respondent even opted to “fill in” for us a short period data gap, run the “figures” through some highly contrived programme and come up with this graph below. WMO “guidelines” were given in support of this concoction stating

“However, the period from 1961 to 1990 has been retained as a standard reference period for long-term climate change assessments.“

Almost inevitably the Talkshop near flat-line graph transformed to this almost meteoric rise shown below.

I am by nature an iconoclast and I do not accept what I view as almost alchemy in setting up artificial and unnecessarily complex rules and regulations on how you should do things to satisfy established institutions requirements. I actually regard this “alternative” data presentation as completely invalid. What it has demonstrated to me, however, is why that particular reference period may have been chosen which was not selected until the early 2000s when political motivations had become well established in institutions such as the Met Office.

Leaving aside the effect of a known cooling period from the later 1940s through to mid 1970s, the 1961 to 1990 period was largely lacking in any significant automation of sites. Those known as “climate only” stations (i.e. not synoptic forecasting units) were much more prevalent and recording only once daily. As is demonstrated here with Keele (and I will be highlighting others soon) even those sites offering multiple daily readings prior were effectively “dumbed down” for alleged uniformity of comparisons. That this degraded reading’s quality and introduced over recording of cold events was probably well understood by the meteorologists but it made actually regularly getting readings from volunteer observers more likely. Once daily only was far less arduous than multiple observation times – some readings were better than no readings. This is very similar to frequent poor station siting for ease of observers rather than accuracy of observations.

I assert that climate science, either by design or ineptitude, is deliberately using false base comparisons (1961 to 1990 with spurious figures) more akin to sharp sales practices than genuine objective science to creatively account for rising temperatures. This must be exposed and stopped.

Source: https://tallbloke.wordpress.com/2025/11/29/keele-dcnn4167-the-issue-of-twice-daily-readings-being-ignored-and-a-startling-revelation/