“Spike Analysis: 1 year on.” Detailed research into Met office instrumentation and siting standards.

I make no apology for this post being nothing more than a re-post of the excellent detailed work by Dr Eric Huxter. This level of pains-taking work deserves as much publicity as possible. It is self explanatory and, though very detailed, makes readily understandable points. I strongly recommend following Eric’s research and his blog.

“I have been looking in detail at the pattern of UK Daily Maximum Temperatures since 26^th April 2025, therefore one year on what has this shown?

The UK Meteorological Office (MO) report the ‘Daily Weather Extremes’ on their webpage:

UK daily weather extremes – Met Office

And through their ‘X’ feed.:

Met Office (@metoffice) / X

from which the previous day’s National Maximum (and Minimum) Temperature can be collected, along with the weather station at which is was recorded.

Weatherobs.com gives access to 7 days of weather readings from many of the stations in the MO’s Synoptic & Climate Network and it is therefore possible to capture those data for the weather station(s) of interest.

Weatherobs

However these data are only hourly, capturing the 1 minute averages on the hour (or other time intervals according to the station) while the MO continuously records 1 minute averages (the minimum time interval recommended by the World Meteorological Organisation (WMO) from the continuously recording Platinum Resistance Thermometers (PRTs),. The WMO recommend that PRTs should be aspirated ie have a fan assisted airflow through the screen, a practice not adopted by the MO.

These ‘extremes’ are ephemeral data, only available for 24 hours from publication on the webpage but still available within the ‘X’ feed for a while. The time of maximum can only be accessed for a short period after publication of the extremes. As far as I am aware there is no public access consolidated database of the extremes and the timings are not available directly to the public.

PRTs are a totally different measurement technology to the traditional Liquid in Glass Thermometers (LIGTs) for which the Stevenson Screen became the standard housing.  Ray Sanders, continuing his excellent work on weather stations, has clearly explained the limitations.

Stevenson Screens – ” The aesthetic and moral codes of the suburban domestic garden.” An explanatory report. | Tallbloke’s Talkshop

Their response times to changes in temperature is short, therefore they will register changes that the LIGTs would not, creating a noisy signal with clear spikes, giving Maxima well above that the LIGTs would have recorded.

The existence of temperature spikes is easily demonstrated from the deviation of the Maximum Minute Temperature (MMT) compared to the Maximum Hourly Temperature (MHT) but how significant are these spikes?

The foundation of this study is the establishment of a baseline with which to compare daily readings. As already reported in previous posts this baseline was derived from a purchased year’s minute data from the MO for a CIMO1 site (Rothamsted). Two Probability Function Distributions were derived: 1) Deviations of temperature minute by minute in the next 60 minutes from On The Hour Minute Averages 2) Minute to minute deviations. For any day’s National Maximum, and other stations where the Daily Maximum is known, the probability of a given spike can be calculated. Where the time is known the spike is Maximum Previous Hour Temperature (MPHT) – MMT and when the exact time is not known it is MMT-MHT.

Where the exact time is known the change per minute can be calculated which can be compared to the minute to minute PDF, although this is of limited value since it is an average, but on occasions it can illustrate the spike tendency.

Over the year the results show that significant spikes are a common feature of the MO’s reported temperature extremes.

From the data collected it is clear that the weather stations featuring in the MO announcements of extremes on many occasions do not represent anything other than the conditions in the immediate locality, yet are trumpeted as of national significance.

From Chi²  the probability of the Observed distribution of temperatures above the previous/maximum hour being the same as the Expected Rothamsted distribution is <0.001% both seasonally and for the year.

As with much of our world view the temperature narrative is governed by the metrology and we are now far better at measuring at shorter intervals and more synoptically. The PRT technology was introduced in the UK from 1990 and progressively, over 25 years, the weather stations have become automated, using PRTs and even manual stations use PRTs for their measurements. The UK Temperature record therefore is the product of 2 eras, the LIGT Era (1884 to 1969, 106 years covering recovery from the Little Ice Age into the Industrial Era and the mid 20^th century cooling) and the current PRT Era (1990 to date). In the PRT Era the UK temperature record from 1990 shows a step change in the maximum temperature value,

the cyclical trend strength being amplified eg 10, 20, 30 year trends:

the climate signal (30 year mean) becoming more pronounced:

More National Daily Records being set with a greater increase in the magnitude of the record:

Of the 366 National Daily Maximum Temperature Rcords set between 1858 & 2026 there is a Mean of 2 per weather station and 19.1% Stations set 51.1% Records. From May 2022 26.6% Stations have recorded 77.0% of Maxima, a self selecting group of stations where the maximum is the result of dynamic local factors, rather than the regional temperature.

The spikes are a result of the interplay of a number of dynamic factors which do not produce a consistent outcome and which are therefore not able to be accounted for by homogenising the national data but are recognised in the CIMO rating (which seems to be regarded as irrelevant by the MO and its cheerleaders). The MO try to get round differences in siting by the use of ‘Anomalies’, which are not anomalies in the true scientific sense since they are simple deviations from a given baseline. To be an anomaly they should be at least 3 standard deviations away from the baseline mean, but this step is ignored by Climate Science who treat all deviations as de facto anomalies. This lack of context ie standardisation, is compounded each time the data are averaged ie daily, regional and national,and outliers ie these common spikes, can have a huge impact on the final value, which is regarded as ‘true’. .

The dynamic factors are a function of  the PRT technology in inadequately ventilated Stevenson Screens (Aitken effect with, as already referenced in earlier posts, Minimum Assumed Ventilation being achieved <99.9% of the time) set within inappropriate sites (CIMO 3/4/5) where windbreaks further reduce the %MAV, artificial surfaces, create localised heat sources, and anthropogenic heat is released.  THE CIMO rating indicates the underlying scale of this effect with additional error (for maxima) of +1°C CIMO 3, +2°C CIMO 4 and +5°C CIMO 5.The effect is inconsistent and unpredictable but nevertheless exists and it provides a stochastic boost to the recorded Maximum Temperature at these stations. The importance of local factors has been illustrated in previous posts eg Pershores, Gosport Fleetlands.

From 26/04/2025 data has been obtained from 112 Weather Stations with 437 Temperature Maxima (365 UK Daily Maximum) averaging 3.9 per station Of these readings 275 (62.9%) have a known time (to the minute GMT) that Maximum Temperature was recorded, 161 (57.7%) occurring before the published maximum hourly reading The mean temperature spike above the Previous Hour is 1.21°C  with a 7.5% probability of occurring at a CIMO 1 site. For all readings the average spike above Maximum/Previous Hour is 0.70°C with a 18.8% probability of occurring at a CIMO1 site. Given that 62.0% happen on/before Maximum hour the average spike  for all dates without a known time taken from hour before the maximum is 1.7°C with a 2.9% probability  Mean ventilation is 36.7% of minimum assumed for PRTs. Only 58 (60.4%) of these stations have wind data.

The 62 spikes with greater than 3 Standard Deviations (true anomalies) are:

The most prone stations are (anomalous mean spikes):

The anomalies in the average change per minute treated as minute to minute change:

Another illustration of the impact of these dynamic impacts on temperature recoding is the ability of the MO to successfully forecast Regional Maximum Temperature 16 hours in advance. ‘X’ publishes the forecast for UK regions at 23:00. Of these predictions to date the MO success rate for Daily Maximum Temperatures (From 05/2022) is 9.1% correct ie Daily Maxima within +/-0.5°C). From 01/2025 58.9% have been in the Region that had been forecast. 88.0% of the readings come from stations rated CIMO 3/4/5 (greater than their representation in the Synoptic & Climate Network) of which 24.30% had <30 years record and 50.3% were established in the prt era. mean error from reporting stations is +2.28° average anomaly (in mo sense) of national daily maximum +2.1°c above forecast for uk +2.88°c region which it was recorded. to date run cimo 3>

The MO is wedded to the popular Anthropogenic Global Warming narrative, or Climate Change as it is now known. The view that Climate Change is the independent variable, driving the weather in the UK is patently false, as Climate is the 30 year mean of weather data (WMO), but it does demonstrate the problem of modellers gaining the keys of the Climate Science kingdom (as warned against by HH Lamb, the founder of Climate Science), as the models often work by reversing causality. Tails wagging dogs is the default for Climate Models unfortunately. The choice of CO₂ as Occam’s Razor to cut through the Gordian Knot of Climate Complexity leaves an awfully large number of loose ends, which are frequently ignored. The change in metrology is one such loose end but the MO knows it is getting warmer and the PRTs confirm this, so where is the harm in the maximum temperature data being noise rather than signal? As already refernced in earlier posts research has shown that in the conversion to PRTs 3% of the data demonstrated >1 °C difference from LIGTs. This is enough to bias the national record from the start. Even so the fact that 97% had up to 1 °C difference is also worrying, given that the MO often reports temperature on weatherobs.com to 1 hundredth of a degree. The propensity of PRTs to spike , as demonstrated by the data from the past year of publically available data, reinforces the bias and should be taken into account. The public facing MO uses maximum daily temperature data that produces a mirage of rapid temperature change, feeding project fear of ‘Climate Change’, whereas the mirage is purely the result of the heat haze produced by spikey PRTs and the PRT Premium.”

Source: https://tallbloke.wordpress.com/2026/04/28/spike-analysis-1-year-on-detailed-research-into-met-office-instrumentation-and-siting-standards/