NOAA’s Updated 2024 Global Average Temperature Anomaly Trend Continues to Decline Indicating a Weaking El Nino and No “Climate Emergency”

Not sure what you mean by that. I’m not claiming any sort of proof from that simple model. Just responding to the claim that there is zero evidence of CO2 causing warming.

Try using a dictionary. Proof and evidence are different things.

I’ve asked you before what you would consider evidence that I creasing CO2 caused warming. So far crickets. It’s obvious that no evidence will ever be acceptable to you. You will just deny any link as an article of faith.

Why? I’m quite prepared to accept I’ve made a mistake, but would prefer it if you explained what in your opinion makes it “physically meaningless” rather than just accept it on your say so.

I take it you can’t or won’t say why you think it’s physically meaningless.

“uncertainty in the anomalies is at least 1.5C and probably more like 15C”

Talk about physically impossible models.

Amazing. Karli will say I’m arrogent for expecting him to justify his vague assertions. Yet the demands I answer him whilst he uses these childish insults.

The question is why do I think it’s impossible that the world could be 15°C warmer last year than it was in the 1950s. This is based on the assumption we are using the GUM definition of uncertainty, a value that describes a range of values in which it is reasonable to attribute a value to the measurement.

Fair enough. I can’t prove that’s physically impossible, just as no one can explain why a sensitivity of 2.4°C is physically impossible.but it seems very unlikely. For a start a change if 15 over a year or a century would require some very strong forcing. I suspect it would be an extinction level event. If the data is to be believed annusl averages have not changed by more than a degree or so over the last century. The ice ages were not 15°C colder than today.

If there had been a year in recent times that was that warm or cold I would have expected it to not pass without comment. Given that I don’t think it’s reasonable to believe this year was 15°C warmer it colder thanaby other yearon the last few hundred. And hence my pont a out the GUM definition, and the use of the word “reasonable”. An interval of ±15°C is just not reasonable even if you had no confidence in the thermometer readings.

Secondly, it’s an absurdity to think that 1000s of instrument readings could contrive to give a global average that was 15°C out. It would be pretty incredible for there to be many daily readings that had that big an error. For the annual global temperature to be out by that much would require every single thermometer in the world to have that size if error every day for a year, and for all the errors to be in the same direction. No thing’s impossible but it’s so improbable as to be virtually impossible. And even if it did happen you would have to assume that somebody would notice.

Of course, what karlo et al want to do is ignore the GUM definition of uncertainty, with all it’s reliance on statistics and probability, and instead have invented the concept of uncertainty as a “zone of ignorance”, or The Great Unkown. This is a magical place where the usual rules of probability no longer apply and anything that can possibly happen will happen, or at least is as likely to happen as anything else.

Even then I’m not sure how you get to 15°C of uncertainty. Pat uses ignorance to claim that the correct uncertainty is the worst case. The average of a million readings each with an ignorance zone of ±2°C will have an uncertainty of ±2°C. Because it’s always possible that every reading will be at the same edge of the interval, and as long as we worship ignorance any attempt to refine that interval is heretical. But even Pat only keeps his ignorance to around ±2°C. It will take an overabundance of ignorance to get that to ±15°C.

If you didn’t want an answer you shouldn’t have demanded am answers. As expected you ignored it, and obviously avoided the part where I addressed your non probability based fantasy.

“You want to discuss statistics, what is the possible standard deviation of a distribution with a range from +90 to -50?”

Why do you keep asking me to do this work for you, whilst constantly claiming I’m mathematically illiterate?

OK – I’ll use some old BEST gridded data, as that’s gives an estimate of the absolute temperature for each grid point. I’ll use 2022 as that’s the last complete year I have, and I’ll get the average annual temperature for each grid point. The standard deviation is 14.3°C.

Now I’m going to assume that you are still under the delusion that the SD of temperatures is the uncertainty of the mean. That is clearly nonsense. If it were you would expect to see fluctuations in the annual averages of at least ±28°C. What you have here is the uncertainty, if you based your annual average on one randomly chosen location each year.

“Don’t tell me that anomalies have their own uncertainty based on their small values.”

You really don’t like being told the truth do you? Of course anomalies have a smaller distribution across the world. The standard deviation of annual anomalies in 2022 was 0.74°C. That by most peoples standards is a smaller number that 14.3.

“It’s not a matter of physically possible.”

So why even mention it? What use is an uncertainty if it doesn’t describe thing thing that are physically possible?

“It’s a matter that the uncertainty grows.”

That’s your problem. You know I don’t accept it. You know you have not been able to offer any evidence either mathematically literate or experimental. Why do you think endlessly claiming it as if it was true, will be in any way convincing.

And repeating the same nonsense a dozen times in the same comment is even less convincing. If you want people to believe that the uncertainty of the mean is the same as the uncertain to of the sum – you need to actually offer some evidence that doesn’t depend on you being unable to follow a simple equation.

How many more times does it need to be explained, I am not trying to forecast the future?

Nor am I trying to create a circulation model, or any thing like that. It’s simply looking at what the correlation is between multiple variables. The point which you keep dancing round is that it does demonstrate that you can explain most of the warming using CO2. That doesn’t prove it’s the cause, but it does give the lie to claims that there is no correlation.

“Correlation without some hypothesis…”

We have a hypothesis.



The world and every one in it. The hypothesis is available to every one. It’s a lie to claim the hypothesis doesn’t exist.

Really? You would prefer it if your GP assumed all people will react identically to the same drug?

But that’s not a functional relationship. A functional relationship means that the same input will always give the same output.

A deterministic (or functional) relationship is an exact relationship between the predictor x and the response y.

A statistical relationship, on the other hand, is not an exact relationship. It is instead a relationship in which “trend” exists between the predictor x and the response y , but there is also some “scatter.”

https://online.stat.psu.edu/stat415/book/export/html/875

Whether Jim actually means a functional relationship, or is just misusing the term, is another question. I’ve asked him many times what he means by it, but have never got a coherent answer.

You keep complaining that I don’t post references, but when I do you ignore it and just assert your definition is correct.

I suspect you assume “functional” on this context means “how something works” which is an easy mistake to make given the word has overloaded meanings. But im the context of a functional relationship it means “like a function” and not “working, practical, useful” or any of the othe synonyms.

And none of this would be s problem if you didn’t keep throwing these terms into your comments without understanding their meaning. It just ends up being a massive distraction from the point you think you are making.

I assume that what you actually mean is that in your opinion any hypothesis must be defining s physical process. That’s where I’m disagreeing with you. Hypothesis can be tentative speculations, based on some observed phenomenon. You do not need an explanation as to why there is a correlation for the hypothesis to be valid. You might not try to understand why it happens before you’ve tested if it does happen.

And in the case of CO2 it’s not relevant, because the “why” was around long before the evidence that it does happen. The prediction that rising CO2 levels woulf rise temperatures is around long before it could be tested.

“The issue is whether or not you know all the factors involved in the functional relationship.”

You are never going to know all the factors. And more to the point, you do not need to know all the factors to have a hypothesis. The hypothesis may just be that there exists a correlation between two factors.

From your own quote “A hypothesis is a tentative statement about the relationship between two or more variables.”

You are insisting that all hypothesis have to describe a complete system with exact predictions. But it’s usually only once you’ve done the experiments that you can begin to describe the detailed model. From the same Wiki page

When a possible correlation or similar relation between phenomena is investigated, such as whether a proposed remedy is effective in treating a disease, the hypothesis that a relation exists cannot be examined the same way one might examine a proposed new law of nature. In such an investigation, if the tested remedy shows no effect in a few cases, these do not necessarily falsify the hypothesis. Instead, statistical tests are used to determine how likely it is that the overall effect would be observed if the hypothesized relation does not exist. If that likelihood is sufficiently small (e.g., less than 1%), the existence of a relation may be assumed. Otherwise, any observed effect may be due to pure chance.

“That does *NOT* mean the functional relationship doesn’t exist!”

your claim was you couldn’t have a hypothesis without defining the functional relationship.

“That’s a STATISTICAL hypothesis”

and that’s the basis of much of science.

Ye Gods and Little Fishes, he’s still wittering on about this. Quite amazing the lengths people will go to avoid accepting something that doesn’t agree with their religious beliefs. Just impossible to accept that a correlation between CO2 and temperature is evidence that the greenhouse effect is real. So instead we have hand waving about it being physically impossible, ever more inflated claims about uncertainty – currently up to 15°C. And then deep dives into epistemology, to try to erase the idea that there is even an hypothesis.

So let’s trawl through this again.

“You keep forgetting that the hypothesis must be testable to be legitimate. ”

It is testable. I tested it, remember, that’s why you are so agitated. It did not fail the test.

“I can hypothesize that evil angels are causing the warming.”

Why, so can I, or so can any man; But will they come when you do call for them?

If you can demonstrate a statistically significant correlation between evil angles and global temperatures, I’ll listen. But first you will have to demonstrate how you measured them.

“Malarky! CO2 has been higher when we were in a glacial period! ”

Strange how quickly you ignore uncertainty when it suits you. I assume you are talking about the Ordovician period, over 400 million years ago. The “experiment” I’m talking about was conducted over a period when most other conditions were more or less constant. Go back 400 million years and you also need to take into account all the major differences, especially the the fact that solar output was somewhat less.

““you never prove a hypothesis”
Tell this to Newton. Tell this to Gauss. Tell this to Ohm.”

OK, I will when I see them next, but I think Newton may object to the claim that he proposed hypothesis.

I take it you are not a fan of Popper. Neither am I, but I still think that it’s impossible to prove anything in an absolute sense. Best you can say it’s it’s probably true, or almost certainly true, but that requires at some level statistical reasoning.

“Statistical tests can only show correlation.”

Not really true, and by the same token, non-statistical tests might only show correlation.

“If CO2 and temp is not correlated then they aren’t correlated.”

You just keep proving you know nothing about statistics.

“ Take Ohm’s law, E = IR. That functional relationship was thought to be true for decades”

Which is why you should make a distinction between thinking something is true, and proving it is true.

You don’t make any mention if the assumptions in Ohm’s law. The functional relationships ly hold for ideal conditionssuch as no change in temperature.

“It’s not just a binary choice, “true” or “false”. Add in “incomplete”. “

Personally I prefer to think of it as not true or false, but a spectrum of probabilities. Nothing can be proven to be true or false but we can day the probability approaches 0 or 1. Although I think almost 1 is a more difficult proposition. It’s usually close to one given any number of assumptions assumptions.

“I prove it to be true with every single use. ”

Again, you really need to define what you mean by “prove”.

We are probably talking at cross-purposes again. I take prove to mean you have shown that something is certainly true, not just for the circumstances you’ve tested, but for all possibilities. It’s impossible to prove that is the case empirically, becasue you can’t test all possible values, and even if you could, it only shows it only worked when you tested it, not that it will always work.

But if you take “prove” to mean it seems to work every time for normal uses, and it’s never let me down. Then yes you can say it’s a proven principle. It’s just not what I would take to mean a scientific proof.

I’m not sure if Ohm’s law is a good example in any case – Newtonian gravity is a better example of a law that can be shown to be correct across most practical circumstances, but is still ends up being dis-proven as a universal law.

It is not a universal law, and that’s nothing to do with the “quantum level”.

Here’s a comment that sums it up:

Ohm’s Law is not a construct which can be derived. It is essentially a generalized observation. It is only useful for a few materials (conductors and medium resistivity), and even then virtually all of those materials show deviations from the ideal, such as temperature coefficients and breakdown voltage limits.

Rather, Ohm’s Law is an idealization of the observed behavior of these materials. As the saying goes, “All models are wrong. Some models are useful.” In this case, Ohm’s Law is extraordinarily useful, but that doesn’t make it universal. Semiconductors, for instance, do not follow Ohm’s Law in any large sense, and look how widespread their use is.

https://physics.stackexchange.com/questions/195016/how-can-one-derive-ohms-law

“I didn’t say this. You are making up another strawman to argue with.”

Huh? I’m saying what I take “prove” to mean. I’m asking you to explain what you take it to mean. I’m still waiting.

“I never said this. Stop putting words in my mouth. Stop creating strawmen to argue with.”

I was not putting words in your mouth. I was giving my opinion of what I thought you were trying to say. Hence starting with “I think your general point here…”. And I’m not going to be lectured on strawmen from someone who continuously claims I’m saying all uncertainties are random and Gaussian, when I’ve repeatedly explained why I’m not saying that.

“That means that *something* other than CO2 is a major factor – which climate science totally ignores.”

Nobody ignores El Niños. There is nothing remotely surprising that over a short period, temperatures cool down after an El Niño.

Your double standards is so obvious here. You reject out of hand 140 years of data showing a statistically significant correlation between CO2 and temperature, just claiming it’s meaningless and claiming uncertainty is ±15. But then claim a 7 year period proves something.

You’re memory is so bad. You’ve repeated this nonsense numerous time and I keep explaing why you are wrong.

Time is not part of the model
I know because I wrote it. The model it Anomaly ~ log(CO2) + ONI + AMO +AOD + TSI.

Time is not used. The only way time comes into it at all is I use a lag of 1 year on each variable. I doubt if that makes much of a difference.

The fact that I then show the predicted values on a time series graph does not mean time is being used. It ‘s just easier to understand then if the plot was in a random order.

This like pulling teeth. Time is not a variable in the model. At no point does it multiply the year by a constant. It’s used as an index simply to ensure that the response on temperature ion one year is based on the independent variable states from he previous year. That does not make time an independent variable. There is no assumption that things will change linearly over time. If the states of the independent variables are the same in 1999 as they were in 1899, the model would predict the same average value in 2000 and 1900.

This is really pathetic. You make a mistake and when I correct you, you just accuse me of lying. Time is not an independent variable in the model. I know because I wrote it. If you don’t believe me, try it yourself.

For the record here’s the summary

 Family: gaussian 
  Links: mu = identity; sigma = identity 
Formula: GISS | mi(ci95/2) ~ CO2t + ONI2t + AODt + AMOt + TSIt 
   Data: cli.df (Number of observations: 141) 
  Draws: 8 chains, each with iter = 5000; warmup = 2500; thin = 1;
         total post-warmup draws = 20000

Regression Coefficients:
          Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
Intercept   -49.20     25.98  -100.40     1.97 1.00    33735    17085
CO2t          2.42      0.05     2.32     2.52 1.00    32503    18625
ONI2t         0.08      0.01     0.06     0.10 1.00    31422    15543
AODt          0.08      0.02     0.04     0.13 1.00    31013    17269
AMOt          0.21      0.04     0.12     0.29 1.00    30878    15474
TSIt          0.02      0.02    -0.02     0.06 1.00    33829    16841

Further Distributional Parameters:
      Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sigma     0.07      0.01     0.06     0.08 1.00    16501    14495

Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).

Time is simply not a variable. I just drew time series graph with the prediction for each year as a convenient way of showing what the output looks like.

I could have given you the graph with CO2 as the x-axis. No time in the graph at all.

Or against TSI

Or you could answer my question. If you are claiming that Frank is saying the model is physically impossible just because of claimed massive uncertainty interval that just demonstrates your lack of understanding of what the words mean.

How does thinking there is a couple of degrees of uncertainty in each annual anomaly result in the model being physically impossible?

The fact that he seems unwilling to actually explain his comments, and lives it to the likes of you to interpret his assertions says it all as fat as I’m concerned.

And as always, the hypocrisy of this is fully evident. A single simple model showing a possible link between CO2 and temperature is dismissed as physically impossible due to supposed uncertainties, yet not a peep about alm the claimed pause, or claims that 7 years of CRN data prove that all warming is caused by EL Nińos.

It’s alright. I didn’t expect an answer.

“You refuse to understand that you are using ONLY the stated values of the anomalies to create your trend.”

Not really true. In my later attempt I included random measurement error in the anomaly values. You of course won’t like that as it’s assuming the errors are Gaussian and random.

“That is physically impossible since the stated values do *NOT* represent true values. “

uou really don’t understand how any of this works. No model will ever be exact. There will always be simplifications. That doesn’t make the model “physically impossible”.

“Pat has long ago abandoned trying to educate you on uncertainty…”

Ha, good one. I needed a laugh.

His education has mostly been him refusing to answer any question, suggesting I was part if a mathematical mafia, and refusing to accept the simplebfact that negative standard deviations cannot be negative.

“you can’t just create ONE, and only one, trend line based on”

Your attempts to “educate” me are hampered by your inability to ever address what I’m saying. I keep telling you there is not one known true trend. Every coefficient in my model has an associated uncertainty. You really need to stop inventing straw men and try to engage with what’s actually being said. And you need to accept the possibility that you are wrong on some points. If you don’t you will always be in your cargo cult – the easiest person to fool yourself.

“Throw away everything you know about statistics”

That seems to be the Gorman way – yes.

“Measurements don’t deal in sampling groups, sampling a population, confidence intervals, or any other statistical analysis.”

TG: “None so blind as they who will not see.”

“Measurements result from observations.”

What do you think statistics results from?

“There is no sampling size calculation to obtain a low SEM. That is why the GUM says:”

Firstly – nobody said anything about the SEM. We are talking about Expanded uncertainty and what karl called uncertainty intervals.

Secondly – nobody cares that the GUM gets triggered by the phrase standard error of the mean. It makes no difference what you call it – the statistical basis is the same.

“The primary purpose of statistical parameters in measurement uncertainty is to DEFINE an internationally accepted definition of the intervals to use in describing the dispersion of observations values that can be attributed to the measurand.”

Why do you never actually quote the definition in the GUM. It is not the dispersion of “observations values” that can be attributed to the measurand. You are just describing what Tim insists was abandoned 50 years ago. True value + error. What uncertainty actually describes is the level of uncertainty of your observation, and in GUM terms this is dispersion of values that could reasonably be attributed tot eh measurand.

“Whereas a Type A standard uncertainty is obtained by taking the square root of the statistically evaluated variance“

Gosh, you mean the standard deviation is the square root of the variances. If only you hadn’t thrown away everything you knew about statistics, maybe that wouldn’t have been such a revelation.

“That is the internationally accepted definition of the interval to be used when declaring standard uncertainty.”

It is not an interval. You never get that. The standard uncertainty is equivalent to the standard deviation. You can use it to define an interval, based on the expanded uncertainty. But the standard uncertainty is just a positive number.

“The standard deviation of the mean can only be useful under strict conditions, primarily observing the EXACT SAME THING.”

And you are back to making things up. You can take a mean of different things, that mean will come from a distribution that can be estimated, and can be characterized by the standard error of the mean, or if your prefer the standard deviation of the mean.

If you use your own definition of uncertainty, then that is what you need if you want to describe the distribution of expected observations of the mean. If your standard deviation of the mean is 1, you expect around 95% of all future sample means to be with ±2 of the population mean.

But as always, if you don’t want to treat the mean as a measurand – stop using the concepts of measuremnt uncertainty on it.