• This is a political forum that is non-biased/non-partisan and treats every person's position on topics equally. This debate forum is not aligned to any political party. In today's politics, many ideas are split between and even within all the political parties. Often we find ourselves agreeing on one platform but some topics break our mold. We are here to discuss them in a civil political debate. If this is your first visit to our political forums, be sure to check out the RULES. Registering for debate politics is necessary before posting. Register today to participate - it's free!

An Article from WUWT

LowDown

LowDown

Curmudgeon
DP Veteran
Joined
Jul 19, 2012
Messages
14,185
Reaction score
8,768
Location
Houston
Gender
Male
Political Leaning
Libertarian
Lots of complaining about links to wattsupwiththat.com from certain people. The main complaint seems to be that links to WUWT are posted at all. They appear to want restrictions on free speech here to the effect that only blogs they approve of can be linked.

In any case, here's another link to WUWT. We'll see how our alarmist friends can do with this thesis, which you can only read about in blogs and the like. It hasn't been possible to get it published in peer review literature for reasons that seem pretty obvious to me. We'd like to hear about views of the science; appeals to authority do not impress because we think the authorities are whack, compromised, corrupted.

Needle in a Haystack. Global Climate Model Error

In short: While general circulation models, i.e., global climate models, agree with each other, that is, they have good precision, this scientist shows that it's impossible for them to be accurate, that is, to actually predict future temperatures or conditions, because of errors that get propagated through the model. Specifically, errors in assumptions for cloudiness get passed on from iteration to iteration of the model. The errors don't cancel out because they are not random, they are systemic. Instead, the errors from the previous iteration get added to that of the next iteration.

I always suspected that this would be a problem with the models, but I didn't have the skill with statistics to express it formally. When you use an iterative model, when the predicted state at one time is used to calculate the next state, then the errors are bound to build up.
 
I'm already tired of the whining about alternative news. Zero-fund correctly saw the corruption in Goldman-Sachs because the MSM was too busy slurping after big banks.

It was the freaking National Inquirer that broke the story about John Edwards extra-marital baby. The MSM actually had the story but was sitting on the information.

It was a blogger that showed Hillary's fainting spell while the MSM was putting up a wall of protection about Hillary's health issues and denounced anyone who dared question her health.

The MSM was fully and completely complicit in allowing the DNC to cheat against Bernie Sanders and in fact helped to facilitate that action and just went whole hog in it's endorsements of HIllary Clinton.

I've learned my lesson from now on. ALL NEWS SOURCES are legitimate. I used to buy into the kool-aid of "good sources" but not anymore. Al-Jazeera? Alex Jones? WND? Slate? RT? Huffington Post? Random Blogger online? It's all good to me from now on.

**** the NYT. The WashingtonPost can go straight to hell. I can't believe I actually used to respect MSNBC's coverage. Forget about it CNN. CNN were actually worse than MSNBC when it comes to being in the tank for Hillary which is saying something because MSNBC markets itself as a liberal haven.
 
LowDown said:
Lots of complaining about links to wattsupwiththat.com from certain people. The main complaint seems to be that links to WUWT are posted at all. They appear to want restrictions on free speech here to the effect that only blogs they approve of can be linked.

In any case, here's another link to WUWT. We'll see how our alarmist friends can do with this thesis, which you can only read about in blogs and the like. It hasn't been possible to get it published in peer review literature for reasons that seem pretty obvious to me. We'd like to hear about views of the science; appeals to authority do not impress because we think the authorities are whack, compromised, corrupted.

Needle in a Haystack. Global Climate Model Error

In short: While general circulation models, i.e., global climate models, agree with each other, that is, they have good precision, this scientist shows that it's impossible for them to be accurate, that is, to actually predict future temperatures or conditions, because of errors that get propagated through the model. Specifically, errors in assumptions for cloudiness get passed on from iteration to iteration of the model. The errors don't cancel out because they are not random, they are systemic. Instead, the errors from the previous iteration get added to that of the next iteration.

I always suspected that this would be a problem with the models, but I didn't have the skill with statistics to express it formally. When you use an iterative model, when the predicted state at one time is used to calculate the next state, then the errors are bound to build up.
Click to expand...

Models end results are how they are programmed.

It only takes one bad variable algorithm to make a model bad.

Models all based on the indoctrinated way of thinking, will show similar results.
 
Jack Hays said:
[h=2]Climate modelers open up their black boxes to scrutiny[/h][FONT="][FONT=inherit]Posted on[/FONT] [URL="https://judithcurry.com/2016/11/05/climate-modelers-open-up-their-black-boxes-to-scrutiny/"]November 5, 2016[/URL] | 219 comments[/FONT]
by Judith Curry
Paul Voosen has written a remarkable article in Science about climate model tuning.
Continue reading →
Click to expand...

Greetings, Jack. :2wave:

This sounds like something that is long overdue, IMO. :thumbs: Most people don't just go along with those who say "because we said so," especially if it means it's going to cost them more money! Glad to finally see more transparency, but time will tell. :shock:
 
polgara said:
Greetings, Jack. :2wave:

This sounds like something that is long overdue, IMO. :thumbs: Most people don't just go along with those who say "because we said so," especially if it means it's going to cost them more money! Glad to finally see more transparency, but time will tell. :shock:
Click to expand...

Greetings, Polgara.:2wave:

I assume you've got an eye on OSU vs UM.
 
Jack Hays said:
Greetings, Polgara.:2wave:

I assume you've got an eye on OSU vs UM.
Click to expand...

YOU BETCHA! This has always been "the game of the year" for both teams - OSU vs That team up North! :lamo
 
Jack Hays said:
[h=2]Climate modelers open up their black boxes to scrutiny[/h][FONT="][FONT=inherit]Posted on[/FONT] [URL="https://judithcurry.com/2016/11/05/climate-modelers-open-up-their-black-boxes-to-scrutiny/"]November 5, 2016[/URL] | 219 comments[/FONT]
by Judith Curry
Paul Voosen has written a remarkable article in Science about climate model tuning.
Continue reading →
Click to expand...


They can get their models to match data from the past because they know the precise observations about cloud forcing, aerosols, and so on. The point of the OP is that these can't be predicted ahead of time and so they will be a source of systemic error going forward, and that error accumulates with time. The error quickly becomes so large that it renders the models useless for predicting the future. It's impossible.

Funny thing is that all of this is known. Weather forecasting models incorporate this knowledge. It sets a limit on how far into the future that weather models will remain accurate, typically only 5 days. Apparently, climate modelers have never consulted either professional statisticians or experts in forecasting. The difference is that weather forecasters are accountable for their accuracy and so far climate modelers have not been.
 
LowDown said:
They can get their models to match data from the past because they know the precise observations about cloud forcing, aerosols, and so on. The point of the OP is that these can't be predicted ahead of time and so they will be a source of systemic error going forward, and that error accumulates with time. The error quickly becomes so large that it renders the models useless for predicting the future. It's impossible.

Funny thing is that all of this is known. Weather forecasting models incorporate this knowledge. It sets a limit on how far into the future that weather models will remain accurate, typically only 5 days. Apparently, climate modelers have never consulted either professional statisticians or experts in forecasting. The difference is that weather forecasters are accountable for their accuracy and so far climate modelers have not been.
Click to expand...
You make a very good point! like the hurricane forecast, the uncertainty is larger the further out they get.
Storm_track.jpg
 
LowDown said:
Lots of complaining about links to wattsupwiththat.com from certain people. The main complaint seems to be that links to WUWT are posted at all. They appear to want restrictions on free speech here to the effect that only blogs they approve of can be linked.

In any case, here's another link to WUWT. We'll see how our alarmist friends can do with this thesis, which you can only read about in blogs and the like. It hasn't been possible to get it published in peer review literature for reasons that seem pretty obvious to me. We'd like to hear about views of the science; appeals to authority do not impress because we think the authorities are whack, compromised, corrupted.

Needle in a Haystack. Global Climate Model Error

In short: While general circulation models, i.e., global climate models, agree with each other, that is, they have good precision, this scientist shows that it's impossible for them to be accurate, that is, to actually predict future temperatures or conditions, because of errors that get propagated through the model. Specifically, errors in assumptions for cloudiness get passed on from iteration to iteration of the model. The errors don't cancel out because they are not random, they are systemic. Instead, the errors from the previous iteration get added to that of the next iteration.

I always suspected that this would be a problem with the models, but I didn't have the skill with statistics to express it formally. When you use an iterative model, when the predicted state at one time is used to calculate the next state, then the errors are bound to build up.
Click to expand...

LowDown said:
People do get banned for being nasty and repeatedly violating policy. At RealClimate all that is needed for banning is a dissenting opinion no matter how nicely and respectfully stated.

Go to WUWT and you will find plenty of dissent. Now go find me dissent at RealClimate.
Click to expand...
Let's knock over 2 of your false claims with one link. I noticed a poster on that rehash of a rehash of a Pat Frank video on WUWT gave a link to RealClimate that shows Pat Franks ideas are not only wrong but incredibly naive.

What the IPCC models really say « RealClimate

Read all the comments (there are pages of them). Plenty of 'dissenters' posting on Real Climate, and they get a polite discussion of the science from scientists and other posters.
 
Quaestio said:
Let's knock over 2 of your false claims with one link. I noticed a poster on that rehash of a rehash of a Pat Frank video on WUWT gave a link to RealClimate that shows Pat Franks ideas are not only wrong but incredibly naive.

What the IPCC models really say « RealClimate

Read all the comments (there are pages of them). Plenty of 'dissenters' posting on Real Climate, and they get a polite discussion of the science from scientists and other posters.
Click to expand...

No, this is not "dissent" on RealClimate. It's true believers arguing about details. There's no real dissent there at all. Show me a comment that actually consists of dissent there.

Is Pat Frank himself allowed to comment on RealClimate? I don't see a comment from him anywhere there, but he commented extensively at WUWT. I conclude that he was excluded from the discussion at RealClimate. No doubt a multitude of other skeptics are excluded. How can one trust any discussion there?

The response from RealClimate to Frank's argument is pretty simple. They insist that any error in the models is random and does not accumulate from iteration to iteration but rather is canceled out. But this can't be true. If they put an erroneous value for cloud forcing into the model then that error will add from one iteration to the next. The error will accumulate. What Frank defines is not the likely error but the range of possible error that the model could produce from this arrangement. It shows that bad values for cloud forcing could result in huge errors from the model. (I can't believe that this point is controversial.)

Again, the people who run short term models for weather forecasting know all about this. Small errors in the starting state of the weather or small inaccuracies in the calculations will add up as the model steps through each time point rendering predictions of the weather out past 5 or 10 days very inaccurate.

Are you people capable of thinking through these issues for yourselves? If so then demonstrate that ability now.
 
reinoe said:
I'm already tired of the whining about alternative news. Zero-fund correctly saw the corruption in Goldman-Sachs because the MSM was too busy slurping after big banks.

It was the freaking National Inquirer that broke the story about John Edwards extra-marital baby. The MSM actually had the story but was sitting on the information.

It was a blogger that showed Hillary's fainting spell while the MSM was putting up a wall of protection about Hillary's health issues and denounced anyone who dared question her health.

The MSM was fully and completely complicit in allowing the DNC to cheat against Bernie Sanders and in fact helped to facilitate that action and just went whole hog in it's endorsements of HIllary Clinton.

I've learned my lesson from now on. ALL NEWS SOURCES are legitimate. I used to buy into the kool-aid of "good sources" but not anymore. Al-Jazeera? Alex Jones? WND? Slate? RT? Huffington Post? Random Blogger online? It's all good to me from now on.

**** the NYT. The WashingtonPost can go straight to hell. I can't believe I actually used to respect MSNBC's coverage. Forget about it CNN. CNN were actually worse than MSNBC when it comes to being in the tank for Hillary which is saying something because MSNBC markets itself as a liberal haven.
Click to expand...

Nice examples.

I agree, most MSM, especially the NYT are no more credible than some random blogger.
At least the random blogger is passionate and honest about what they write.
I think this 2016 election has shown us the dark recesses of the MSM and how horribly they suck.

Again, though, to their credit, the NYT at least admitted it.
Not that they are going to change anything or their lean at all.
But at least they admitted to their extreme bias in print.
I will credit them for that.
None of the others have.
never will, either.
 
LowDown said:
No, this is not "dissent" on RealClimate. It's true believers arguing about details. There's no real dissent there at all. Show me a comment that actually consists of dissent there.

Is Pat Frank himself allowed to comment on RealClimate? I don't see a comment from him anywhere there, but he commented extensively at WUWT. I conclude that he was excluded from the discussion at RealClimate. No doubt a multitude of other skeptics are excluded. How can one trust any discussion there?

The response from RealClimate to Frank's argument is pretty simple. They insist that any error in the models is random and does not accumulate from iteration to iteration but rather is canceled out. But this can't be true. If they put an erroneous value for cloud forcing into the model then that error will add from one iteration to the next. The error will accumulate. What Frank defines is not the likely error but the range of possible error that the model could produce from this arrangement. It shows that bad values for cloud forcing could result in huge errors from the model. (I can't believe that this point is controversial.)

Again, the people who run short term models for weather forecasting know all about this. Small errors in the starting state of the weather or small inaccuracies in the calculations will add up as the model steps through each time point rendering predictions of the weather out past 5 or 10 days very inaccurate.

Are you people capable of thinking through these issues for yourselves? If so then demonstrate that ability now.
Click to expand...

Amazing how you managed to read 473 comments (many of them quite long and technical) in...less than 30 minutes.Didn't see the posts from Pat Frank himself or any of the other 'dissenters' like Gerald Browning? Shows you didn't bother to read the comments. Are you incapable of clicking "Next"?

So, doesn't read the comments, doesn't understand GCMs, doesn't understand that GCM's are bounded by physics, doesn't understand that Frank's so-called 'model' was a simple linear model, and continues to makes false claims. Sounds like a typical ideologically motivated, scientifically illiterate, anti-science WUWT worshipper who ignores what scientists who have expertise in their fields say, but totally accepts whatever some scientist with no expertise in climate sciences says, is unaware of the the huge body of science on this topic, and thinks they are 'thinking for themselves' but just repeats silly pseudoscience conspiracy nonsense from WUWT.
 
Last edited:
Quaestio said:
Amazing how you managed to read 473 comments (many of them quite long and technical) in...less than 30 minutes.Didn't see the posts from Pat Frank himself or any of the other 'dissenters' like Gerald Browning? Shows you didn't bother to read the comments. Are you incapable of clicking "Next"?

So, doesn't read the comments, doesn't understand GCMs, doesn't understand that GCM's are bounded by physics, doesn't understand that Frank's so-called 'model' was a simple linear model, and continues to makes false claims. Sounds like a typical ideologically motivated, scientifically illiterate, anti-science WUWT worshipper who ignores what scientists who have expertise in their fields say, but totally accepts whatever some scientist with no expertise in climate sciences says, is unaware of the the huge body of science on this topic, and thinks they are 'thinking for themselves' but just repeats silly pseudoscience conspiracy nonsense from WUWT.
Click to expand...

The issue with model uncertainty is very simple. Say you have a clock with a defective mechanism such that it runs 1 minute slow every day. After 30 days it will be 30 minutes off. Then say you want to create a mathematical model that predicts the uncertainty of such defective clocks. The range of uncertainty of a defective clock is going to depend on the particular defect the clock has, running fast or running slow, and how big the error is. Your model is going to have to account for the fact that the daily error is going to add up in time, whether it is positive or negative. It will not be the case that the clock will be one minute slow one day and one minute fast the next, so that the errors cancel out. Each particular clock will always be slow or fast, and the errors of previous days will add to that of the next day. The mathematical expression of this is root mean squared (RMS), so that regardless of the sign of the error the total uncertainty is always positive. This is systemic error.

What the climate modelers are insisting at RealClimate and WUWT is that errors in models are not systemic. They say they are random, and they cancel out with time. This can't be true. If the cloud forcing parameter is slightly off from the true mean then it will produce an error that will always be the same sign in each iteration of the model, and that will accumulate. There is no way to eliminate this type of error, so the models will always be too inaccurate to be useful.

https://www.physics.umd.edu/courses/Phys276/Hill/Information/Notes/ErrorAnalysis.html
 
Last edited:
LowDown said:
Lots of complaining about links to wattsupwiththat.com from certain people. The main complaint seems to be that links to WUWT are posted at all. They appear to want restrictions on free speech here to the effect that only blogs they approve of can be linked.

In any case, here's another link to WUWT. We'll see how our alarmist friends can do with this thesis, which you can only read about in blogs and the like. It hasn't been possible to get it published in peer review literature for reasons that seem pretty obvious to me. We'd like to hear about views of the science; appeals to authority do not impress because we think the authorities are whack, compromised, corrupted.

Needle in a Haystack. Global Climate Model Error

In short: While general circulation models, i.e., global climate models, agree with each other, that is, they have good precision, this scientist shows that it's impossible for them to be accurate, that is, to actually predict future temperatures or conditions, because of errors that get propagated through the model. Specifically, errors in assumptions for cloudiness get passed on from iteration to iteration of the model. The errors don't cancel out because they are not random, they are systemic. Instead, the errors from the previous iteration get added to that of the next iteration.

I always suspected that this would be a problem with the models, but I didn't have the skill with statistics to express it formally. When you use an iterative model, when the predicted state at one time is used to calculate the next state, then the errors are bound to build up.
Click to expand...

These arguments simply make logical sense. Random errors tend not to effect a models projections because it is assumed that random errors occur in equal frequency both + and - and even themselves out. If a model consistently errors from observation in one direction, in this case + then it can be safely assumed that the model has a systemic bias that will be cumulatively more wrong as you run the model into the future or the past.

This can be seen in the famous spaghetti graph that has been used by both sides for years. Where the various models are tuned to known data they appear to all be modeling the same thing, but when left to hindcast or forecast solely on the power of the model they all diverge as their systemic flaws accumulate.
 
Last edited:
Damn!! Didn't any of you denialists actually watch the video?

Frank clearly states that the errors are both positive and negative. So these errors are most likely random and not systemic. And then when Frank adds up all the possible error and comes up with his huge error ranges that are completely out of the realm of possibility he is really just making up more denialist BS.

It is no wonder that his paper has been rejected by 6 scientific journals. And those were just the ones who replied to his submissions. It would be interesting to know how many didn't bother to reply.
 
Buzz said:
Damn!! Didn't any of you denialists actually watch the video?

Frank clearly states that the errors are both positive and negative. So these errors are most likely random and not systemic. And then when Frank adds up all the possible error and comes up with his huge error ranges that are completely out of the realm of possibility he is really just making up more denialist BS.

It is no wonder that his paper has been rejected by 6 scientific journals. And those were just the ones who replied to his submissions. It would be interesting to know how many didn't bother to reply.
Click to expand...

Thank you for an excellent example of confirmation bias.
 
Jack Hays said:
Thank you for an excellent example of confirmation bias.
Click to expand...

Says the guy who spams this forum with almost nothing but biased articles. Oh... the irony!! :shock::3oops::lol:

Hey!! Why don't you actually join the discussion and tell me exactly what I said that shows confirmation bias? Both LowDown and jmotivator admitted that if the errors are both positive and negative then they would cancel out.

If anyone is committing confirmation bias it is all the denialists here who are forgetting that Frank said "plus or minus".
 
Buzz said:
Damn!! Didn't any of you denialists actually watch the video?

Frank clearly states that the errors are both positive and negative. So these errors are most likely random and not systemic. And then when Frank adds up all the possible error and comes up with his huge error ranges that are completely out of the realm of possibility he is really just making up more denialist BS.

It is no wonder that his paper has been rejected by 6 scientific journals. And those were just the ones who replied to his submissions. It would be interesting to know how many didn't bother to reply.
Click to expand...

Don't you comprehend?

Errors are actually not likely to cancel each other out. No matter what they say. The probability is actually low.

Try running the statistics of the probability of cancellation.
 
LowDown said:
The issue with model uncertainty is very simple. Say you have a clock with a defective mechanism such that it runs 1 minute slow every day. After 30 days it will be 30 minutes off. Then say you want to create a mathematical model that predicts the uncertainty of such defective clocks. The range of uncertainty of a defective clock is going to depend on the particular defect the clock has, running fast or running slow, and how big the error is. Your model is going to have to account for the fact that the daily error is going to add up in time, whether it is positive or negative. It will not be the case that the clock will be one minute slow one day and one minute fast the next, so that the errors cancel out. Each particular clock will always be slow or fast, and the errors of previous days will add to that of the next day. The mathematical expression of this is root mean squared (RMS), so that regardless of the sign of the error the total uncertainty is always positive. This is systemic error.

What the climate modelers are insisting at RealClimate and WUWT is that errors in models are not systemic. They say they are random, and they cancel out with time. This can't be true. If the cloud forcing parameter is slightly off from the true mean then it will produce an error that will always be the same sign in each iteration of the model, and that will accumulate. There is no way to eliminate this type of error, so the models will always be too inaccurate to be useful.

https://www.physics.umd.edu/courses/Phys276/Hill/Information/Notes/ErrorAnalysis.html
Click to expand...

I finally watched the video.

Great stuff, and in my opinion, spot on as to what is probably happening in the climate models.
 
Buzz said:
Damn!! Didn't any of you denialists actually watch the video?

Frank clearly states that the errors are both positive and negative. So these errors are most likely random and not systemic. And then when Frank adds up all the possible error and comes up with his huge error ranges that are completely out of the realm of possibility he is really just making up more denialist BS.

It is no wonder that his paper has been rejected by 6 scientific journals. And those were just the ones who replied to his submissions. It would be interesting to know how many didn't bother to reply.
Click to expand...

Apparently you didn't understand the video, or didn't watch it.

He showed high correlation for a systemic error.

Go back to about 18:43 and see the lag run error correlations.
 
Lord of Planar said:
I finally watched the video.

Great stuff, and in my opinion, spot on as to what is probably happening in the climate models.
Click to expand...

Yet again your complete ignorance about climate models shows that your amateur opinions are irrelevant when it comes to science.
 
LowDown said:
The issue with model uncertainty is very simple. Say you have a clock with a defective mechanism such that it runs 1 minute slow every day. After 30 days it will be 30 minutes off. Then say you want to create a mathematical model that predicts the uncertainty of such defective clocks. The range of uncertainty of a defective clock is going to depend on the particular defect the clock has, running fast or running slow, and how big the error is. Your model is going to have to account for the fact that the daily error is going to add up in time, whether it is positive or negative. It will not be the case that the clock will be one minute slow one day and one minute fast the next, so that the errors cancel out. Each particular clock will always be slow or fast, and the errors of previous days will add to that of the next day. The mathematical expression of this is root mean squared (RMS), so that regardless of the sign of the error the total uncertainty is always positive. This is systemic error.

What the climate modelers are insisting at RealClimate and WUWT is that errors in models are not systemic. They say they are random, and they cancel out with time. This can't be true. If the cloud forcing parameter is slightly off from the true mean then it will produce an error that will always be the same sign in each iteration of the model, and that will accumulate. There is no way to eliminate this type of error, so the models will always be too inaccurate to be useful.

https://www.physics.umd.edu/courses/Phys276/Hill/Information/Notes/ErrorAnalysis.html
Click to expand...

You show you don't have a clue how climate models work. And you still haven't read the comments in the RealClimate thread, including the lesson in basic math and statistics that Gavin had to give Pat Frank (who had about a dozen long-winded posts there despite your lies). Your credibility and integrity is non-existent
 
Jack Hays said:
Thank you for an excellent example of confirmation bias.
Click to expand...

Says the guy who floods this subforum with gazillions of click-bait links to pseudoscience conspiracy blogs.
 
You must log in or register to reply here.
Back
Top Bottom