Climate Models for the Layman

[longview]

There are actually twenty shown there, though best estimates are provided for only nineteen of them. The ten for which publication details are actually shown on the chart are the ones which were new since AR4: Out of those, 4 provide an ECS best estimate lower than 2.4 degrees and 5 provide a higher estimate (three of them above 3 degrees by the looks, though I'm just eyeballing it this time), while Scharwtz 2012 is the one for which no best estimate is provided.

Clearly, it is not the case that "most" studies support an estimate on the low end of the IPCC range. On the contrary, most studies support a 'best estimate' around the 3 degree mark +/-20% (with many of them suggesting 5-95% confidence ranges that extend above 4 degrees), and your decision to blindly dismiss this scientific research as "just based on speculation" purely because you have been proven wrong in your assertions is disappointing, to say the least.

As Poor Debater has suggested, even simply comparing ln(CO2) with recorded temperatures implies a climate sensitivity somewhere in the 2.2 to 3.8 degree range (if memory serves in the past he has pegged his personal estimate at 2.7 or 2.8 degrees).

Fair enough, I was just looking at the named labels, I wonder what the dashed lined are indicating?
As to most, the IPCC did not include all the studies, I was glad that they included Lindzen at below 1C, because
the climate is clearly more complex than the models show.
As to the ln(CO2), that only counts the one variable, there is more happening.

 

[Lord of Planar]

All of it.

No models, just data:
View attachment 67214354

This stupid graph again?

Remember the lagged solar graph I did, using your lagged methodology with a better fit? I had over 0.9!

 

[Lord of Planar]

Now you're just being silly.

Just?

Isn't that normal?

 

[Lord of Planar]

Please explain why my graph is misleading, or withdraw your slanderous accusation.

LOL...

Seriously?

I explained it long ago. You can take any two sets of data with a similar trend, and show a high correlation between them. Even if they aren't related at all.

So...

Are you being intellectually dishonest, or ignorant to the facts of science?

 

[Mithrae]

This stupid graph again?

Remember the lagged solar graph I did, using your lagged methodology with a better fit? I had over 0.9!

I don't. What did you use, some ridiculous 50 year lag?


Edit: Unlike CO2 forcing, solar influence regularly increases and decreases: The relatively small peak value we had in 2015 is not still influencing global temperatures, because the sun is now cooler. By contrast the CO2 levels from 2015 are still influencing the climate, because that CO2 is still in the atmosphere (plus some more on top) and there hasn't yet been enough time to reach equilibrium.

I got 0.514 correlation for the period 1880 to 2013 against an estimate of the varying solar influence on temperatures - an estimate which is pretty much identical to your own. So if you managed to produce a much higher level of correlation, it must be because you decided to abandon everything you've learned in favour of cheap deception.

It would be interesting to see the comparison if we had more reliable information about 18th and early 19th century temperatures though.

Spoiler

 

[Mithrae]

Fair enough, I was just looking at the named labels, I wonder what the dashed lined are indicating?
As to most, the IPCC did not include all the studies, I was glad that they included Lindzen at below 1C, because
the climate is clearly more complex than the models show.
As to the ln(CO2), that only counts the one variable, there is more happening.

Of course there's more happening - if there weren't, all the data points would fall exactly on the trendline and the correlation would be 1 :lol: We know that CO2 has been one of (if not the) major drivers of temperature over the past century. Of particular note, we know that its influence has been most pronounced in the past 50 years or so, since that's when concentrations have risen most rapidly, and that's exactly what we see in Poor Debater's graph - more noisiness at the lower CO2 concentrations and a better fit at the higher end:

 

[flogger]

We know that CO2 has been one of (if not the) major drivers of temperature over the past century.

This is pure nonsense given we still have no idea what the correct climate sensitivity of CO 2 is. It is currently pure guesswork. There are many factors we cannot as yet quantify any one of which may be more or less significant. The inability to model clouds and water vapour for instance (which represents 95% of all greenhouse gasses) means that the guesswork involved here is nearly total so these claims of the primacy of CO2 are pure wishful thinking

Of particular note, we know that its influence has been most pronounced in the past 50 years or so, since that's when concentrations have risen most rapidly, and that's exactly what we see in Poor Debater's graph - more noisiness at the lower CO2 concentrations and a better fit at the higher end:

Correlation is not causation. We have had dozens of such warming phases like the one we are seeing today even over the last 10 millenia and they have been for the most part beneficial to humanity.



The added CO2 has also led to an increased greening of the planet too (especially in the more arid regions) which cannot be a bad thing surely ?

https://www.csiro.au/en/News/News-releases/2013/Deserts-greening-from-rising-CO2

The wealth tranference you seek is not going to be achieved with this agenda giving its fading importance with both the media and the public over the last decade so perhaps you should think up something else



Its is interesting to note too that the poor countries that are claimed to be most affected by climate change are also the ones least concerned about it.

Go figure

MYWorld2015 Analytics

 

[Mithrae]

This is pure nonsense...

When you're asserting falsehoods about water vapour's influence, posting a graph from Greenland as if it represented global temperatures, promoting one attribution of observed changes based on modeled responses to CO2 (global greening) whilst blindly denying others, spewing random nonsense about "wealth transference" and appealing to popular interest as a barometer for scientific truth, it's a pretty good bet that everyone will ignore your views regardless of where they stand on the issue :roll:

 

[flogger]

When you're asserting falsehoods about water vapour's influence, posting a graph from Greenland as if it represented global temperatures, promoting one attribution of observed changes based on modeled responses to CO2 (global greening) whilst blindly denying others, spewing random nonsense about "wealth transference" and appealing to popular interest as a barometer for scientific truth, it's a pretty good bet that everyone will ignore your views regardless of where they stand on the issue :roll:

But I cornered you on the wealth transference issue (that was your ultimate motivation for promoting this here) years ago.

The plain fact is with ever declining public support this is never going to happen for you however much of you peddle your subjective interpretations of the latest junk science

 

[Lord of Planar]

I don't. What did you use, some ridiculous 50 year lag?


Edit: Unlike CO2 forcing, solar influence regularly increases and decreases: The relatively small peak value we had in 2015 is not still influencing global temperatures, because the sun is now cooler. By contrast the CO2 levels from 2015 are still influencing the climate, because that CO2 is still in the atmosphere (plus some more on top) and there hasn't yet been enough time to reach equilibrium.

I got 0.514 correlation for the period 1880 to 2013 against an estimate of the varying solar influence on temperatures - an estimate which is pretty much identical to your own. So if you managed to produce a much higher level of correlation, it must be because you decided to abandon everything you've learned in favour of cheap deception.

It would be interesting to see the comparison if we had more reliable information about 18th and early 19th century temperatures though.

Spoiler

View attachment 67214490

I used the same method he used. Not my graph. I did it just to prove the point of similar trends, and yes. I used a longer lag.

Why would 50 years be ridiculous? Just because you don't like it?

His 21 years for CO2 is a joke too.

Like I said, to prove a point, that two similar trends will have a close correlation, even if not actually related.

 

[Mithrae]

I used the same method he used. Not my graph. I did it just to prove the point of similar trends, and yes. I used a longer lag.

Why would 50 years be ridiculous? Just because you don't like it?

His 21 years for CO2 is a joke too.

Like I said, to prove a point, that two similar trends will have a close correlation, even if not actually related.

I told you why it's ridiculous: Solar activity from 1959 is not still contributing to warming 50 years later, or even one year later. On annual or decadal timeframes solar activity fluctuates, while CO2 accumulates. 21 years may or may not be too long, even for CO2, but as I showed the results are similar even at 2 years' lag (r^2=0.864 and implied climate sensitivity around 2.6 degrees).

As for solar values, even ignoring the absurdity of applying lag to its influence, I have serious, serious doubts about your claim to have found a >0.9 match-up in the first place, considering the r-squared value is only 0.25 even with 50 years' lag on solar values (0.271 with 30 years, 0.163 with 70 years).

 

[Lord of Planar]

I told you why it's ridiculous: Solar activity from 1959 is not still contributing to warming 50 years later, or even one year later. On annual or decadal timeframes solar activity fluctuates, while CO2 accumulates. 21 years may or may not be too long, even for CO2, but as I showed the results are similar even at 2 years' lag (r^2=0.864 and implied climate sensitivity around 2.6 degrees).

As for solar values, even ignoring the absurdity of applying lag to its influence, I have serious, serious doubts about your claim to have found a >0.9 match-up in the first place, considering the r-squared value is only 0.25 even with 50 years' lag on solar values (0.271 with 30 years, 0.163 with 70 years).

Yes, it is ridiculous. That's what I am pointing out!

 

[Lord of Planar]

How about reading my posts again, removing your confirmation bias, and keep in mind I was pointing out correlation does not indicate causation.

 

[Mithrae]

How about reading my posts again, removing your confirmation bias, and keep in mind I was pointing out correlation does not indicate causation.

What, are you suggesting that there's no causation between CO2 concentrations and temperature? Or between solar activity and temperature?

You haven't shown a comparable level of correlation between solar activity and temperature to begin with - as I said, even with 30 or 50 or 70 years' lag r-squared remains well and truly below 0.3 - so as far as counterexamples go this one would have be considered pretty piss-poor, don't you think?

So all we've really seen is your ignorance of relatively basic scientific concepts. I'm sorry if you don't like that.

 

[Lord of Planar]

What, are you suggesting that there's no causation between CO2 concentrations and temperature? Or between solar activity and temperature?

You haven't shown a comparable level of correlation between solar activity and temperature to begin with - as I said, even with 30 or 50 or 70 years' lag r-squared remains well and truly below 0.3 - so as far as counterexamples go this one would have be considered pretty piss-poor, don't you think?

So all we've really seen is your ignorance of relatively basic scientific concepts. I'm sorry if you don't like that.

I am simply saying that using correlation doesn't prove squat. There are too many variables for one thing.

His graph is a joke, and that's why I don't flaunt such graphs as valid science.

 

[Mithrae]

I am simply saying that using correlation doesn't prove squat. There are too many variables for one thing.

His graph is a joke, and that's why I don't flaunt such graphs as valid science.

No, you just flaunt different amateur guestimations as being better than real science :lol:

Unless there were some other variable affecting global temperatures precisely in line with ln(CO2), or close to it, those other variables' influence on temperature would (and do) decrease the level of correlation with ln(CO2): For example 1998 and 2016, hot years with strong El Ninos, lie well above the trend line while 1992/93 and 1999/2000, cool years influenced by factors like La Nina or the Pinatubo eruption, lie below it. The lower CO2 regions of the graph appear to be noisier than the higher concentrations, as CO2 became an increasingly significant influence on temperatures.

As far as I can imagine, the only obvious candidates likely to correlate closely even with CO2 - let alone ln(CO2) - are other anthropogenic influences related to population growth: Other anthropogenic GHGs might enhance the perceived trend, while anthropogenic aerosols might depress it. Even those are fairly big "ifs," and blindly speculating that there must be something else entirely which shares the same correlation really is a bit too much like wishful thinking for my tastes. Of course, I am always open to new information.

 

[Lord of Planar]

Temperature lagged six years, but what did I reference it to?

I could have found data that fits better, but this is my first attempt of correlating data outside of the climate sciences.

 

[Lord of Planar]

No, you just flaunt different amateur guestimations as being better than real science :lol:

You have no traits of a scientist with that type of prejudiced thinking.

Bye.

 

[Poor Debater]

I believe I've finally got 'round to replicating your results:

https://docs.google.com/spreadsheets/d/1ofbw72DDspPtzXIRoYhasqy6zF_u55NakdvKeI0qtw8/edit?usp=sharing

But assuming I've done it right, I'm not sure how useful this would be in estimating climate sensitivity, because the results seem to depend very heavily on both the temperature data used and the lag estimate applied.

Comparing ln(CO2) against HadCRUT temperatures lagged 2 years, climate sensitivity is estimated at around 2.2 degrees (r^2=0.845)
Comparing ln(CO2) against GISS temperatures lagged 2 years, climate sensitivity is estimated at around 2.6 degrees (r^2=0.886) (second sheet)
Comparing ln(CO2) against HadCRUT temperatures lagged 21 years, climate sensitivity is estimated at around 3.3 degrees (r^2=0.864) (first sheet)
Comparing ln(CO2) against GISS temperatures lagged 21 years, climate sensitivity is estimated at around 3.8 degrees (r^2=0.897)

For comparison - for any others interested - the r-squared value of a trendline from a regular temperature time series is about 0.76 (third sheet). Using a series starting at 290 (same as CO2) but varying randomly by values between -2 and 2, the average r-squared value was 0.307 (from twenty-five trials). Using a series starting at 290 but randomly increasing by 0 to 2 (hence on average ending up around 426 'ppm,' compared to the actual 404ppm), the average r-squared value was 0.8118 from twenty-five trials (the two highest values were 0.841 and 0.837). From a simple linear increase from 290 to 426ppm (+1 each year) the r-squared value would be 0.814. The actual observed CO2 concentrations provide a closer fit because they have increased more rapidly in recent decades (as temperatures in general have tended to).

View attachment 67214388

Each temperature dataset has a unique lag that results in the highest temp/CO2 correlation; although in every case it is CO2 that leads and temp that lags, which formalizes the direction of causality in a statistical sense. For each dataset, the lag represents the thermal inertia of the oceans, as measured by that dataset. The range of max-correlation lags ranges from 7 to 21 years, with 17 years being the mean. Longer lags also result in higher sensitivities, but the range of sensitivities implied ranges between 2.6 and 3.7°C, right in line with the IPCC canonical range.

Sensitivity can be easily computed by slope * ln(2) for any dataset or lag.

 

[Poor Debater]

This stupid graph again?

Remember the lagged solar graph I did, using your lagged methodology with a better fit? I had over 0.9!

You didn't use my methodolgy. You cooked up your own methodology that you still have not explained to this day.

 

[Lord of Planar]

You didn't use my methodolgy. You cooked up your own methodology that you still have not explained to this day.

No, I used your method with solar and used an unspecified lag. The only difference was my lag was more than 20 years.

Keep in mind, I was showing how ignorant it is to think such a correlation is correct. I most certainly will not contend that simplistic method was correct for solar, just like I contend it is not correct for CO2.

 

[flogger]

You didn't use my methodolgy. You cooked up your own methodology that you still have not explained to this day.

Thats particularly ironic coming from you and your history of doctoring graphs and charts to get them to say what you want them to :roll:

 

[Poor Debater]

LOL...

Seriously?

I explained it long ago. You can take any two sets of data with a similar trend, and show a high correlation between them.

But if there is causality between data, datasets should have a similar trend. You're taking real-world evidence and hand-waving it away because you don't like the political implications.

Even if they aren't related at all.

Seriously? Now you're saying CO2 has no effect on temperature? Is that how deep in denial your brain has to go to ignore evidence you find politically unpalatable?

Are you being intellectually dishonest, or ignorant to the facts of science?

Says the intellectually dishonest guy who ignores science with every post he makes.

 

[Lord of Planar]

Wrong again.

Too bad. Remain ignorant.

 

[Jack Hays]

More on Curry’s climate model study saying they are ‘not useful as projections for how the 21st century will actually evolve.’

Guest essay by Larry Hamlin The new climate model study by Dr. Curry addressed in the February 21 WUWT article provides some very powerful conclusions regarding the unsuitability of using climate models for purposes of projecting future global climate behavior. In the Executive Summary of this study Dr. Curry delivers the bottom line on the…

. . . Figure 6 from Dr. Curry’s new study shows an updated IPCC AR5 Fig. 11.25 demonstrating that using temperature observations through 2015 the CMIP5 climate model simulations project warming, on average, about a factor of 2 higher than observed temperatures. She attributes this excessive projected warming to “a combination of inadequate simulations of natural internal variability and oversensitivity of the models to increasing carbon dioxide (ECS).”

Figures 2 and 3 from Dr. Curry’s new study show global surface temperature anomalies since 1850 and climate model simulations for the same period respectively. She notes that the modeled global temperatures match closely for the period 1970 to 2000 but fail to capture the warming period between 1910 and 1940.

The failure of climate models to address the earlier warming period is described by Dr. Curry as follows:
“If the warming since 1950 was caused by humans, then what caused the warming during the period 1910-1940? The period 1910-1940 comprises about 40% of the warming since 1900, but is associated with only 10% of the carbon dioxide increase since 1900. Clearly, human emissions of greenhouse gases played little role in this early warming. The mid-century period of slight cooling from 1945-1975 – referred as the ‘grand hiatus’ – has also not been satisfactorily explained.”. . .