PRT Spikiness, ongoing site comparisons and a Kew Gardens heat map. How to conduct real research not false attribution.

A second Guest post by Dr Eric Huxter following on from his “Aspiration” research and moves further onto the subject of “spikiness” of readings originally highlighted here. Eric’s Frayedends blog is a must read site – ” The use of CO2 as Occam’s razor to cut through the Gordian knot of climate complexity leaves an awful number of loose ends.”

Probability of UK Daily Maximum Temperature Spikes

The UK Meteorological Office presents the British Public with a barrage of weather information, especially of ‘extremes’, without any context. Their mission seems to be to educate the public to be afraid of the natural variability of weather and to see it as the result of ‘Anthropogenic Climate Change’, thus reversing cause and effect. The language of ‘anomalies’, without any context of Standard Deviations, and the idea that climate change leads to extreme weather, without the context that climate is the dependent variable, suggests a carefully constructed programme of mis-information. The information released to the media seems designed to ‘feed the fear’ and since it come from authority, it is not questioned.

One of the extremes that is used is the UK Daily Maximum Temperature, via the Meteorological Office Extremes Page or ‘X’ and any possible records is avidly seized upon and the public is taught that anything over 25°C is hot and over 30°C is ‘sizzling’, without the context of previous such temperatures over the past 150 years. I have already posted about the spikes in Daily Maximum Temperatures, above the 1 minute averages published via weatherobs.com at hourly intervals. One key issue about these spikes is their probability of occurrence under normal weather conditions ie how well do they represent the ‘true’ meteorological signal.

To investigate this I have taken the data from the Hull University weather station, aspirated with 5 minute averages, which, although a CIMO 4 site, provides a good baseline for ‘natural’ temperature conditions. From these data maximum deviation from the previous hour can be calculated. These figures can classified according to 0.1°C bins and the frequency of occurrence obtained for rising temperatures.

The cumulative frequency graph then allows a probability of a given temperature or greater to be established. The best fit is a 5th order polynomial, with the probabilities at the upper end manually adjusted to create a smooth curve.

The probability of a given spike value or greater is plotted and the observed daily maxima difference from previous/maximum hour superimposed on an arbitrary Y scale.

It should be noted that these spikes maybe underestimates, as for currently only 8 of 87 is the exact time of the actual maximum temperature known. The calculation of the rest assumes that the maximum hourly 1 minute average temperature occurs before the actual daily maximum but in reality it could be a falling temperature and occur after.

This suggests that the majority of the observed spike ie difference between the reported daily maximum at a station and its previous/maximum hour’s temperature, could fall within a natural meteorological signal at a CIMO 4 site. Taking 5% probability as a threshold 7 of the recent 104 maximum daily temperatures fall outside this threshold as individual readings.

However the Daily Maximum Temperature records cluster in number of hot sites, with 41/97 daily maxima from 5/37 weather stations and 62/97 from 12/37 weather stations, 92% of which are CIMO 3/4/5 – over represented as there are 86.6% in the Synoptic & Climate Network.

Since probabilities are multiplicative the actual (compound) probabilities of the observed differences between the reported daily maximum temperature and the previous/maximum hour would be (with a 5% threshold):

It is interesting that Kew Gardens, a supposed CIMO 2 site, used to bolster the temperature recording credibility of nearby also poorly performing Heathrow, performs so badly. Although this is not so surprising if you take into account the amount of hot air vented by surrounding glasshouses and other local heat sources.

The likelihood is therefore that while many temperature spikes may be natural and reflect the ‘true’ meteorological signal of that site with local heat sources (CIMO 3/4/5), a not insignificant minority produce significant spikes as a result of dynamic, very localised heat sources, yet they are part of the Synoptic & Climate record and will contribute to trends and records and while the others maybe consistent with a ‘true’ meteorological signal, the baseline is a CIMO 4 site, albeit aspirated, with an inherent 2°C additional error. It should also be remembered that these not aspirated sites have a systematic tendency to be hotter than an aspirated station.

it would be instructive to run the same analysis on data from a CIMO 1/2 site to obtain the probabilities of given spikes occurring in instruments uncontaminated by local heat sources.

Source: https://tallbloke.wordpress.com/2025/08/02/prt-spikiness-ongoing-site-comparisons-and-a-kew-gardens-heat-map-how-to-conduct-real-research-not-false-attribution/