Hot Air & Large Lies To Go: What HadCET Tells Us
The daily data are in for June and July and Schrödinger’s cat is both in and out of the bag
By way of introduction readers might like to squint back at my articles entitled: Hottest June kills UK fish and threatens insects, June 1976 vs. June 2023, and my five part series entitled: Hot, Hotter, Hottest.
The crew at the Hadley Centre have promptly released the data for July for the Central England Temperature Record (HadCET) – which you may download for your goodselves here.
When it comes to talking about heat, heatwaves, soaring temperatures and global boiling experts do a quirky little thing; they reach for mean daily, weekly or monthly temperatures instead of maximum daily, weekly or monthly temperatures.
Maximum means hot! hot! hot!, whereas the mean means… well… er… an averagey-sort-of-thing. This preoccupation with averages when speaking about temperature extremes doesn’t make much sense until you discover that averages can give you better-looking warming rates, this being a numerical trick TPTB like to play.
What I better do, then, is issue analyses based on daily maxima and analyses based on daily means otherwise alarmists will throw a hissy.
And now for some freshly baked slides….
June 2023
Mean Maximum Daily Temperature
The mean maximum daily temperature reached 22.57°C for June 2023, making it the hottest such temperature within the HadCET dataset since daily records began in 1878.
The silver medal goes to 1976 with a mean maximum of 22.49°C, being just 0.08°C cooler.
Whilst alarmists may consider this an outright ‘win’ what they will have gone and done in their excitement is ignore the fact that these figures are point estimates bathed in sample error.
For some peculiar reason the Met Office and other expertistas forget basic statistical concepts such as sample error so I shall remedy this situation by running an analysis of variance (ANOVA) over the daily mean maximum temperature series for June 1976 and June 2023 (n=60). Herewith the crunch table:
What this is doing is asking if the 0.077°C difference in mean maxima is statistically significant and the answer is a resounding NO! (p=0.947).
In plain English, that scrappy little bit of difference could easily have come about by chance and thus cannot be taken to signify a genuinely warmer 2023.
Shame on anyone suggesting otherwise!
I can rub salt into wounds a little more by running a linear regression (black line) through the June data cloud, which yields a statistically insignificant warming trend of 0.51°C per century (p=0.071).
In addition we discover 10 days at or above 25°C and 2 days at or above 30°C for June 1976, compared to nine days at or above 25°C and zero days at or above 30°C for June 2023.
I shall summarise all this geek-talk by saying June 1976 and June 2023 are highly comparable in their maximum temperature profile for central England, with claims of a hotter June 2023 being based on wishful thinking.
Average Daily Temperature
The mean daily temperature for June 2023 reached 16.96°C , making it the fourth hottest such temperature within the HadCET dataset since daily records began in 1772.
The gold medal goes to 1846 with a daily mean of 18.20°C, being a rather surprising 1.23°C warmer.
As a fair-minded sort of chap I ought to run another of those ANOVAs to see if this difference also passes statistical muster. Here we are…
Computer says no! Even that whacking great difference of 1.23°C fails to reach statistical significance at the 95 percent level of confidence when we consider the daily variation in the mean temperature within the month of June.
That difference could easily have arisen by chance and therefore does NOT represent any hard and fast change that may be ascribed to climate.
N.B. This is one of those moments when we should all stoke our chin thoughtfully and consider how idiotic the Met Office have been in pretending one year is hotter than another because of a fractional change in the recorded temperature that falls well within the bounds of sample error.
To rub the sea salt further we discover nine days at or above 20°C for June 1846, compared to five days at or above 20°C for June 2023.
To rub that salt even further I can reveal a statistically insignificant cooling trend estimated at -0.01°C per century (p=0.909).
June mean temperatures are NOT getting any hotter, folks!
July 2023
Mean Maximum Daily Temperature
Well, here it is – the BIG one – being the time series for mean maximum daily temperature for July for the central England region since records began in 1878.
The eagle-eyed will have spotted a rather sorry-looking orange blob on the right and wonder just what temperature did the hottest July in history manage to scrape.
I shall reveal that July 2023 managed to muster a mean maximum daily temperature of 20.13°C, pushing it down to 77th place. The gold medal went to 2006 (25.60°C), with 2018 taking silver (25.46°C) and 1983 picking up bronze (24.91°C).
There is some good news for alarmists by way of a statistically significant warming trend estimated at a healthy 1.19°C per century (p<0.001).
What is fascinating is that central England Julys have been warming faster than central England Junes, and I suggest this is a koan that we meditate on, for how can one summer month reveal more global warming than the next if global warming arises from global concentrations of atmospheric CO2?
Average Daily Temperature
We see another sorry-looking orange blob on the right and I shall reveal July 2023 attained a mean daily temperature of 16.16°C, this pushing it into 106th place. Whoops.
I wonder if talking heads at UN/WMO and beyond are regretting making all those wild claims after just a few warm(ish) days into July?
Probably not – there’s nothing like a fat salary and promise of a knighthood and/or book deal to stave off the pain of professional embarrassment!
The gold medal went to 2006 (19.82°C), with 1983 taking silver (19.40°C) and 2018 picking up bronze (19.25°C). There is another smidgen of good news for alarmists by way of a statistically significant but rather sedate warming trend estimated at 0.27°C per century (p=0.011).
So there you go. I cheekily asked for bets on July 2023 not making it to the top twenty hottest months of all time believing I was going to be backing a decent horse (judging by Mr & Mrs Dee’s sweater wearing rate) but I didn’t bank on July 2023 coming home with a lame leg at 77th (mean maximum) and 106th (mean) place.
I am hoping this illustrates just how far we can trust our illustrious leaders and the organisations that keep them propped up.
Yes, I’m looking at you BBC, and you Sky News, as well as a rather unethical bunch over at the Met Office who reckon that fractional variation in averaged monthly temperature records is any way indicative of anything robust when it comes to weather.
Yes indeed, we’ve got some long-term warming trends, but we’ve also got some not-so-warming trends and not-really-warming-at-all trends, then we’ve got July warming faster than June under the same skies.
Then we’ve got some cyclical components (check out those fat green LOESS lines) that protagonists love to ignore.
All this requires thinking rather than throwing soup in a psychotic manner.
See more here substack.com
Bold emphasis added
John Dee (not his real name) is a former British government G7-level scientist who now uses his analytical skills to highlight where the public is being lied to on various subjects.
Please Donate Below To Support Our Ongoing Work To Defend The Scientific Method
PRINCIPIA SCIENTIFIC INTERNATIONAL, legally registered in the UK as a company 
incorporated for charitable purposes. Head Office: 27 Old Gloucester Street, London WC1N 3AX.
Trackback from your site.
Alan
| #
It is often said that science is not about producing a theory, rather it is about the experiment or observations need to prove a theory. So what experiment could be done to show that an average temperature is found by first calculating a total and dividing by the number of observations. Try it with two cups of water at 50C. How can a total temperature be found? Putting the two cups of water into a bigger container will give a total of 50C, so is the average 25C? There is no physical quantity equivalent to an average temperature. There would be a law of physics called conservation of temperature for there to be an average temperature. There is of a law of conservation of energy so it makes sense to talk of total energy in the two cups of water, but what would be the point of calculating the average energy of the two cups. That is also meaningless.
The concept of an average temperature is only a guide to weather and deciding when to go on holiday.
Reply
D. Boss
| #
Your reasoning is incorrect. A mathematical quantity like an average is a symbolic construct which corresponds to some physical reality. Math is not god as some believe, but rather a tool to simulate reality. But your reasoning that two cups at 50 degrees C when combined are still at 50C is a direct correspondence to an average! The average is (50 + 50) / 2 = 50! The math does not conclude the average of the two cups is 25C! So your example is mistaken and as such so is your conclusion.
What is the point of an average? Well that can become subjective and can also depart from anything meaningful. You do not decide how to dress for outdoor work based on the expected average temperature, you dress for the expected min and max and duration of each. (example would be a cold morning with frost and 2C warming to 18C by 14:00 and dropping back down to 5C by 19:00 – in that case you would dress in layers so you can moderate your clothing to meet the range of temperatures faced during that 12 hour day spent outdoors – so the average being say 12C is not the correct way to gauge what to wear)
However there are real and correct uses of an average. One example is to make True Power measurements in a pulsing power circuit using a DSO (Digital Storage Oscilloscope) You set up to measure the voltage and current at sufficient sampling rate to yield accurate rise/fall readings, Then you use a math channel to multiply these instant values and obtain an instantaneous power trace. Then you use the cursors to choose an integer number of cycles and take the averagemean of that instantaneous power trace (for an integer number of cycles) and that is the True Power of the circuit/system under test. (you cannot use RMS, or reactive power, or mean voltage times mean current of this kind of waveform – those only work with sinusoidal signals) (i.e. multimeters are erroneous for measuring pulsed, non sinusoidal waveforms) (using the mean voltage times the mean current yields a false result and is an example of when not to use an average or mean)
An average has a meaningful and useful value, but it can also be misleading. Or put another way it can be fit for some purposes and unfit for others.
Reply