Met Office Record Temperature Claims Are Fraudulent

“We’re assessing whether a particular station’s reported temperature is physically consistent with the broader atmospheric state inferred from many independent observations.”

You are measuring DIFFERENT THINGS. The measurement uncertainties of different things ADD, they don’t subtract.

If the uncertainty interval is wider than the difference you are trying to identify, YOU DON’T KNOW IF THE DIFFERENCE IS REAL. You can’t even tell the slope of a trend!

Look at the temperatures in the attached graphic for temps in NE Kansas at 8pm, 7/3/2026. The mean temp is 88F. The standard deviation is 2F for the stated values alone. This assumes the stations are measuring the *same* thing – which of course they aren’t. The actual measurement uncertainty would be the RSS sum of individual components. With 14 components and an assumed measurement uncertainty of +/- 1F the total measurement uncertainty would be +/- 3.7F.

How can you infer the accuracy of any specific measurement station with such a wide range of temperatures, such a wide standard deviation of the data, and the measurement uncertainty of the total?

“No. The measurand is air temperature being measured.”

Uh huh. Each of the temperatures in the image is the same measurand.

Here you are again, trying to lecture people on things you know nothing about.

GUM:
———-
B.2.2
value (of a quantity)
magnitude of a particular quantity generally expressed as a unit of measurement multiplied by a number
EXAMPLE 1 Length of a rod: 5,34 m or 534 cm.
EXAMPLE 2 Mass of a body: 0,152 kg or 152 g.
EXAMPLE 3 Amount of substance of a sample of water (H2O): 0,012 mol or 12 mmol.
—————— (bolding mine)

It is *NOT* the lengths of multiple rods, the mass of multiple bodies, or the amount of substance in different samples.

———————
3.1.4 In many cases, the result of a measurement is determined on the basis of series of observations obtained under repeatability conditions
—————-

Each of those temperatures reflects the measurement of a different measurand, i.e. the value of a property of a different object.

“Who brought up slopes? I didn’t.”

Of course you did! Now you are reduced to saying that a trend line has no slope?

“Why are you treating the observed spread as though it’s primarily measurement uncertainty?”

YOU are the one trying to classify the temperature measurements shown as repeated measurements of the same measurand. That makes the measurement uncertainty into a Type A. In a Type A measurement uncertainty the standard deviation of the observations is the measurement uncertainty.

“That spread is exactly what you’d expect because the stations are in different locations.”

Yep. And that means you ADD the individual measurement uncertainties to get the total measurement uncertainty. The very thing you are trying to say is wrong!

When you add those measurement uncertainty values you quickly reach a point where any differences are subsumed by the measurement uncertainty. You simply don’t know if any differences actually exist or not. Unless you have a crystal ball supplied by climate science.

“Do you really expect expect a station on a hill, another in an urban area, another in a valley, and another over farmland to all report identical temperatures?”

I’m sure you don’t realize it but you just refuted the entire concept of homogenization of temperatures by replacing missing temperature data from a station with the data from surrounding stations.

Homogenization is just one more garbage climate science meme. All it does is spread measurement uncertainty around!

“Most of that spread reflects real spatial variability, not thermometer error. Those are two completely different things.”

Spatial uncertainty is a *sampling* uncertainty, not a measurement uncertainty. Some of the spread is due to MICRO-ENVIRONMENT differences. Different elevations, different geography, different terrain, different pressures, different surroundings such as UHI or evapotranspiration, even the paint condition on each station from UV effects, etc. And, of course, some is due to things like mis-calibration, cloud cover, humidity, etc.

Climate science simply averages all of these temperatures without regard to the accumulation of measurement uncertainty. It’s a result of the garbage climate science meme of “all measurement uncertainty is random, Gaussian, and cancels”.

“Spatial and temporal continuity (or consistency) checks are based on the fact the value or magnitude of nearby observations are usually similar”

I just gave you an image of NE Kansas temperatures that shows this isn’t true. Would comparing Pikes Peak temperature with the temp in Colorado Springs be spatially and temporally of the same value or even magnitude?

In the summer, San Diego and Ramona, CA temps are more than 30F different, yet they are only 30 miles apart. Geography and terrain are totally different. NO SPATIAL SIMALARITY.

If you want, I can get you several examples of temperatures from the east side and west side of a mountain and show you the temporal difference in their profiles. But it would be better if you did it for yourself so you would actually have to learn something on the subject.

I’ve even had climate scientists tell me that homogenization from stations 500 miles away is legitimate because they are in the same time zone!

Values and magnitudes stop being similar at the first hedgerow between stations let alone between a river valley and the associated river plateau.

Take the east/west side of a mountain. Boise vs Fort Collins. Which do you suppose gets the most sun insolation on a diurnal basis? Are their temperature values similar both in value and time?

“The true temperatures are therefore strongly correlated in both space and time.”

So what? The MEASUREMENT of those temperatures is what has uncertainty. Please note that the uncertainty interval has both a plus and a minus value. The *true temperature” can be anywhere in that interval FOR ANY SPECIFIC INDIVIDUAL MEASURMENT VALUE.

That uncertainty interval includes both RANDOM and SYSTEMATIC uncertainty effects. The random effects do *NOT* cancel to zero leaving only SYSTEMATIC uncertainty effects. Having only systematic effects, ones that can be IDENTIFIED, is a requirement in order to validate measurements over time being highly correlated.

You simply don’t know and can’t tell if the individual measurements, even ones taken on consecutive days, are strongly correlated or not when the measurement uncertainty is +/- 5C.

What you are saying, in essence, is that you can know the Great Unknown. That’s the claim of a carnival fortune teller with a cloudy crystal ball. And that’s a perfect descriptor of climate science today!

“#1) Hundreds, if not thousands, of independent temperature measurements from stations across Europe satisfy the JCGM repeatability criterion (see below) because they measure the same measurand: the regional weather”

You apparently don’t even understand what the word “measurand” means in the GUM. It’s a typical example of the knowledge of metrology by climate science.

Here are the requirements for REPEATABILITY:
——————————–
B.2.15
repeatability (of results of measurements)
closeness of the agreement between the results of successive measurements of the same measurand carried
out under the same conditions of measurement
NOTE 1 These conditions are called repeatability conditions.
NOTE 2 Repeatability conditions include:
— the same measurement procedure
— the same observer
— the same measuring instrument, used under the same conditions
— the same location
— repetition over a short period of time.
————————

Note carefully “the same measuring instrument”, “the same location”, and “the same conditions”.

Measurements of “regional weather” does not meet ANY of these requirements.

“I don’t see why a physically consistent body of meteorological evidence (hundreds of weather stations, radiosondes, numerical weather models, knowledge of the evolving synoptic pattern, etc.) wouldn’t qualify as “all available information” when assessing the plausibility of the true value”

“I don’t see” – that’s because like your other compatriots in climate science (bdgwx, bellman, TFN, etc) you ABSOLUTELY REFUSE to study metrology for meaning and context. You cherry pick pieces that you think support your position without regard to whether those “pieces” comport with the meaning and context of metrology.

Each of the hundreds of individual measurements have a measurement uncertainty, that measurement uncertainty represents the Variance of a random variable (i.e. the individual measurement). When you add random variables, each with its own variance, the variance of each individual random variable ADDS. Variance_total = Var1 + Var2 + …

That Var_total becomes the measurement uncertainty for the result of combining all those measurements. If the measurement uncertainty of each individual temperature in the dataset is +/- 0.5C then the uncertainty associated with that data set is (n) * +/- 0.5C. If “n” is 1000 the measurement uncertainty becomes +/- 500C. In other words whatever you do with the dataset is physically useless.

How do you and climate science get around this? By using the unstated assumptions of 1. all systematic uncertainty = 0, 2. all measurement uncertainty is random, Gaussian, and cancels, and 3. the SEM thus becomes the measurement uncertainty which can be made arbitrarily small with more and more individual components being included in the dataset. In other words, GARBABE ASSUMPTIONS. This is even ignoring that the SEM is SAMPLING ERROR, not measurement error.

Then you and climate science refuse to characterize the dataset as a single sample of a lot of components or whether each component represents a single sample of size 1. 1. if it is a single sample then, regardless of the number of entries, the SEM has an in-built uncertainty of 30%, 2. if it is a collection of multiple samples with a sample size of 1 the measurement uncertainty is sum of all of the individual measurement uncertainties.

Climate science makes some more garbage assumptions about the temperature measurements: 1. the measurement uncertainty of each component is Gaussian, 2. that the measurements themselves are not multi-modal, and 3. systematic uncertainty = 0.

From a physical science point of view as well as metrology, climate science should move to using enthalpy and 5-number statistical descriptors. Abandon all the garbage assumptions. We should have data available for at least 40 years to do so.

“Then I have no choice but to believe that you think there was no high pressure over Europe at the time of the heat wave.”

How did you make the leap to this? What you don’t know is the TRUE VALUE of that high pressure. That’s because the measurement uncertainty is a part of the measurement.

YOU SIMPLY CAN’T IGNORE MEASUREMENT UNCERTAINTY.

“Suppose you’re measuring the air temperature at Heathrow Airport at 15:00 UTC on July 1. Now imagine the same thermometer measures that temperature thousands of times over the course of a few seconds. All of those conditions are satisified.”

NO! Not all the conditions are satisfied. 1. Temperature measuring stations do *NOT* make thousands of measurements every second. 2. Because of air flow through the measuring station each of those thousand measurements will be of a different measurand. 3. over one second, hysteresis can cause an asymmetric uncertainty even with digital measuring devices. Meaning the measurement value is dependent on the direction of any temperature change, including over just one second. It’s not even clear if a field measuring device exists with the response time to make a thousand measurements per second.

“LOLLLLLL, you can’t seriously believe that!”

Of course I believe that. It’s the MATH. The issue is that you don’t understand metrology at all!

Again, one more time:

When the measurement uncertainty is greater than the difference trying to be identified, you don’t know if the difference exists or not!

You exceed the measurement uncertainty trigger for just two stations with +/- 1C if you are trying to identify a difference in the tenths or hundreds digit! You don’t need to go any further than that! Trying to add in more stations doesn’t help!

Measurement uncertainty is *NOT* the SEM. The SEM is sampling uncertainty, not measurement uncertainty. You can minimize sampling uncertainty with larger samples. You cannot minimize measurement uncertainty with larger samples of different things.

If you were designing a bridge would you use the SEM of measurements of 100 beams as the measurement uncertainty for shear strength of the beams? Climate science does just that!

“I’m not ignoring measurement uncertainty. I’m rejecting the assumption that every value within a nominal ±5°C interval is equally plausible once you account for the observed synoptic pattern. “

That only means you don’t grasp the concept of measurement uncertainty.

You are using the conclusion as part of the argument.

I know the synoptic pattern because I can ignore the measurement uncertainty involved with it so then I can ignore the measurement uncertainty of the components making up the synoptic pattern.

If you don’t know the actual data values because of measurement uncertainty, then you can’t know the synoptic pattern without measurement uncertainty either.

“Simply absurd. Your interpretation makes the concept of repeatability practically unusable for any continuously varying quantity, including the temperature examples discussed throughout the GUM.”

Talk about absurd. It is perfectly feasible to use multiple sensors to create multiple observations of the same thing at the same time. The issue is the measurement uncertainty associated with each of those sensors contributes to the total measurement uncertainty of their average.

You simply can’t get away from the concept of measurement uncertainty. It’s just plainly obvious that you have never been involved in the field of ACCURATE measuring at all. For instance, how do you handle the effects of inserting a temperature sensor into the flow of a hot gas or liquid that will then cause turbulence in the flow thus affecting the temperature distribution? Inserting multiple sensors will only increase the turbulence in the flow.

What does the measurement uncertainty budget look like for such a situation?

“You’d know that if you read the document as a whole instead of cherry picking isolated passages that you think support your argument. You have a knack for projecting.”

The only people cherry picking around here are you and your compatriots trying to justify how climate science garbage assumptions are “true science”.

You had absolutely *NO* idea of what “repeatability” means when you said temperature data from different measurands using different instruments result in meeting “repeatability” requirements.

Now you are just whining that someone caught you out. The only fix for that is for *YOU* to learn something on the subject before trying to lecture others on how they are wrong.

Only if they meet repeatability requirements! An independent measurement of the temperature in Mexico City at noon today vs an independent measurement of the temperature in Kansas City at noon will tell you almost nothing of use in the physical world.

“Tell him, for example, that the uncertainty of an average decreases as 1/√N under the usual assumptions for independent measurements, “

You are as bad as bellman and bdgwx. They absolutely refuse to identify whether they are speaking of the SEM or the measurement uncertainty when discussing measurements. They just use the word “uncertainty” hoping to be able to use the Equivocation fallacy when called on their assertions being incorrect.

You are doing the same. The “uncertainty of an average” from using 1/sqrt(N) IS SAMPLING UNCERTAINTY. It is *NOT* measurement uncertainty.

My guess is that you will also continue to not specify whether you are talking about sampling uncertainty or measurement uncertainty – just like bellman and bdgwx.

It all stems from the garbage meme of: “all measurement uncertainty is random, Gaussian, and cancels” that is continually found in climate science.