$10.00
Recommend
2 
 Thumb up
 Hide
50 Posts
Prev «  1 , 2  | 

BoardGameGeek» Forums » Everything Else » Religion, Sex, and Politics

Subject: Skeptic finds he now agrees global warming is real rss

Your Tags: Add tags
Popular Tags: [View All]
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
4 
 Thumb up
0.01
 tip
 Hide
  • [+] Dice rolls
Darrin Williams
United States
Allen
Texas
flag msg tools
I have overtext. Yes I do.
badge
I have overtext. How 'bout you !
mbmbmbmbmb
AsgerSG wrote:
. . . just as the Tobacco industry kept contesting healthcare issues connected to smoking at every step of the road,


It's also reminiscent of the way people kept contesting Eugenics a century ago.

There are a lot of great skeptics out there who have put their careers in jeopardy for what they believe.

There's a lot more money in the alarmist business than there is in the skeptic business.

2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Guidance is internal.
United States
Cleveland Heights
Ohio
flag msg tools
mbmbmbmbmb
DarrinWilliams wrote:

There's a lot more money in the alarmist business than there is in the skeptic business.


Are you being sarcastic here or are you referring to being skeptical of industry claims of product safety in the face of scientific evidence to the contrary? Because, I think that there was a LOT more money in selling cigarettes than claiming cigarettes causes lung cancer. Just like I think there is a lot more money in burning oil and coal than there is in measuring sea water temperatures. Maybe that is your point and I didn't get it.
10 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Chris White
United States
Brooklyn
New York
flag msg tools
mbmbmbmbmb
DarrinWilliams wrote:
AsgerSG wrote:
. . . just as the Tobacco industry kept contesting healthcare issues connected to smoking at every step of the road,


It's also reminiscent of the way people kept contesting Eugenics a century ago.

There are a lot of great skeptics out there who have put their careers in jeopardy for what they believe.

There's a lot more money in the alarmist business than there is in the skeptic business.



http://en.wikipedia.org/wiki/Psychological_projection
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
mrspank wrote:
A prominent physicist and skeptic of global warming spent two years trying to find out if mainstream climate scientists were wrong. In the end, he determined they were right: Temperatures really are rising rapidly.

http://news.yahoo.com/skeptic-finds-now-agrees-global-warmin...


Some internal controversy surrounding this study.

http://www.express.co.uk/features/view/280948/Is-global-warm...-

Quote:
Prof Muller appeared on Radio 4’s Today Programme last Friday where he described how BEST’s findings showed that since the Fifties global temperatures had risen by about 1 degree Celsius, a figure which is in line with estimates from Nasa and the Met Office.

When asked whether the rate had stopped over the last 10 years he said they had not. “We see no evidence of it having slowed down,” he replied and a graph issued by the BEST project suggests a continuing and steep increase.

But this last point is one which Prof Curry has furiously rebuttted. In a serious clash of scientific experts Prof Curry has accused Prof Muller of trying to “hide the decline in rates of global warming”.

She says that BEST’s research actually shows that there has been no increase in world temperatures for 13 years.

She has called Prof Muller’s comments “a huge mistake” and has said that she now plans to discuss her future on the project with him. “There is no scientific basis for saying that global warming hasn’t stopped,” she says.

“To say that there is detracts from the credibility of the data, which is very unfortunate.” New research also seems to back up Prof Curry rather than Prof Muller.

A report published by the Global Warming Foundation, which is based on BEST’s findings, includes a graph of world average temperatures over the past 10 years and it is absolutely flat, suggesting that temperatures have remained constant.

This issue is crucial because the levels of carbon dioxide in the air have continued to rise rapidly over the last decade and if temperatures have remained constant during that period it would suggest there is no direct link between carbon gas emissions and global warming.

Previously carbon dioxide emissions – from the burning of fossil fuels and from deforestation – have been considered one of the biggest causes of climate change, the most damaging effects of which are thought to be the melting of the polar ice caps and the rise in sea levels as well as an increase in extreme weather events such as floods and droughts.

“Whatever it is that is going on here it doesn’t look like it’s being dominated by carbon dioxide,” says Prof Curry.

Prof Muller has made it clear that the BEST study was not conducted in order to gauge the causes of global warming, saying the study “made no assessment on how much of this is due to humans and how much is natural”.

He and his scientists – who also included this year’s physics Nobel winner Saul Perlmutter – set out purely to determine once and for all whether climate change had occurred.

The group had been suspicious of previous results which confirmed a rise in global temperatures , believing that their work may have been skewed by the “urban heat island effect” where increasing urbanisation around weather stations was causing the temperature increases recorded over the past 50 years.

But their exhaustive research discovered that the urban heat effect could not explain the global temperature increase of about one degree Celsius since 1950.

IT IS well to point out that Prof Curry is not disputing the one degree Celsius increase. She is disputing Prof Muller’s suggestion that temperatures haven’t levelled off in the last decade.

Indeed she says this global warming standstill since the end of the Nineties – which has been completely unexpected – has wide-reaching consequences for the causes of climate change and has already led many climate scientists to start looking at alternative factors that may have contributed to global warming, other than carbon gas emissions. In particular she has mentioned the influence of clouds, natural temperature cycles and solar radiation.

What she also seems furious about is the way that Prof Muller went about publishing BEST’s results without consulting her and before a proper peer review could be carried out. “It is not how I would have played it,” she has said. “I was informed only when I got a group email. I think they have made errors and I distance myself from what they did. It would have been smart to consult me.”

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
United States
Missouri
flag msg tools
mbmbmbmbmb
Quote:
This issue is crucial because the levels of carbon dioxide in the air have continued to rise rapidly over the last decade and if temperatures have remained constant during that period it would suggest there is no direct link between carbon gas emissions and global warming.


Well, even if it is true that isn't the right conclusion. Likely, it isn't a linear relationshp
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Brian Morris
United States
Raytown
Missouri
flag msg tools
2nd, 6th and 7th Wisconsin, 19th Indiana, 24th Michigan
badge
24th Michigan Monument Gettysburg Pa
mbmbmbmbmb
TheChin! wrote:
DarrinWilliams wrote:

There's a lot more money in the alarmist business than there is in the skeptic business.


Are you being sarcastic here or are you referring to being skeptical of industry claims of product safety in the face of scientific evidence to the contrary? Because, I think that there was a LOT more money in selling cigarettes than claiming cigarettes causes lung cancer. Just like I think there is a lot more money in burning oil and coal than there is in measuring sea water temperatures. Maybe that is your point and I didn't get it.


I think his point is a valid one regardless of global warming. If you say something is wrong there is a lot of money to be made in fixing it. For example if people worry that a nasty flu epidemic is coming there's a lot of money to be made with products that people would purchase to avoid getting sick. On the other hand there isn't a lot of money to be made by saying there will be no epidemic and everything is fine. People in the 50s and 60s made millions building bomb shelters thanks to the Cold War and fear of a nuclear war.

Regardless of the legitimacy of global warming people are making a lot of money off it. Scientists get grants to do research, people buy products because they're environmentally friendly. I use to be in marketing and I tell you the word "green" is a marketing man's dream. I've been out of the business for years now but I suspect that the marketing companies know exactly how much more of a product they can sell just by claiming it's green.

This isn't a knock on global warming. It's just the way the world goes round. Alarmism is good for business.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Sean Chick (Formerly Paul O'Sullivan)
United States
New Orleans
Louisiana
flag msg tools
designer
Fag an bealac! Riam nar druid ar sbarin lann! Erin go Bragh! Remember Ireland and Fontenoy!
badge
Well, I'm afraid it'll have to wait. Whatever it was, I'm sure it was better than my plan to get out of this by pretending to be mad. I mean, who would have noticed another madman round here?
mbmbmbmbmb
I defer to scientists because they are usually right. Their world is based upon the concrete and the measurable. I also defer to them because they know more than me about their given subject. I certainly would not welcome their random commentary on the French Revolution if they pretended it was definitive. But mostly I defer to them because they are usually correct and they show a greater willingness to change their ideas if the data points in another direction. I cannot say the same for historians, politicians, or economic theorists.

What I find funny here is the idea that somehow there is more money to be made as alarmists or that scientists are either in on a conspiracy or just plain wrong. Yet these same people will trot out some Austrian economists as "experts" and not question ideas that cannot be proven scientifically. No sir, capitalism is a "natural" system and the only way to have real capitalism is if it is unregulated. So on economic issues many American conservatives are zealots of the Austrian school but on global warming they become hardcore skeptics. It is a strange sight to behold.
13 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
eknauer wrote:

Some internal controversy surrounding this study.

http://www.express.co.uk/features/view/280948/Is-global-warm...-

Quote:
Prof Muller appeared on Radio 4’s Today Programme last Friday where he described how BEST’s findings showed that since the Fifties global temperatures had risen by about 1 degree Celsius, a figure which is in line with estimates from Nasa and the Met Office.


I was wondering how long it would take you to pick this up.

If you had done a little bit more research, you would have discovered that:

1. Judith Curry has been frantically backpedalling on her blog. She's desperately trying to keep her anti-science following happy without making actually lying about the data, and finding it hard.

2. The graph this controversy was based on comes from the GWPF, an anti-science think tank. They included two months which were present in the BEST data file but not in their graphs, which were calculated from only 47 stations in antarctica rather than 14400 globally. That these points were bogus was obvious to anyone who looked at the data - the uncertainties on the last two months were 30 times greater than the previous month. Removing those points leads to a positive trend.

3. It is impossible to get a statistically significant differentiation between no warming and warming on a 10 year data set. To check this, all you have to do is add an extra year of data and see if it makes a difference. Try it. The trend on 2000-2010.25 is more than double the trend on 2001-2010.25. What Curry and the GWPF were looking at was noise, not signal.

Just to make this absolutely clear:

Anyone who draw a conclusion based on a temperature trend on only a decade (indeed less than about 15 years for the BEST data), is incompetent or dishonest. Let me prove it (h/t Sphaerica):


This image shows the BEST data from 1975, with a series of short term trends plotted on it. Note that every single year since 1975 is part of one or more declining trends. This proves that 2010 is cooler than 1975, that global warming stopped in 1973, 1980, 1987, 1995, 1998 and 2002, and that global warming is a hoax.

Or alternatively, it proves that Curry, the GWPF, and the Mail and Express are lying with statistics, and you were taken in.

Here is a proper analysis from a real statistician. Unfortunately, the statistics of time series analysis required to do this right are generally post-graduate level. So it is understandable that lay observers get taken in all the time.
21 
 Thumb up
0.04
 tip
 Hide
  • [+] Dice rolls
Daniel Edwards
United Kingdom
London
Unspecified
flag msg tools
Pop quiz!

Will Eric:

1. Respond in a reasoned way to the points made in the above post.

2. Go "yeah but" and link to something completely irrelevant.

3. Ignore it completely.

3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
myopia wrote:
Pop quiz!

Will Eric:

1. Respond in a reasoned way to the points made in the above post.



Actually, I think Kevin makes it clear that Curry's response was misinformed.

12 
 Thumb up
0.04
 tip
 Hide
  • [+] Dice rolls
United States
Missouri
flag msg tools
mbmbmbmbmb
Kevin C wrote:

Or alternatively, it proves that Curry, the GWPF, and the Mail and Express are lying with statistics, and you were taken in.

Here is a proper analysis from a real statistician. Unfortunately, the statistics of time series analysis required to do this right are generally post-graduate level. So it is understandable that lay observers get taken in all the time.


As an upfront, I just want to be clear that I am taking issue with what I see as dubious statistics in the link you gave -- subject to just as much lying as Curry (I am not taking issue with the subject of warming, just how statistics are applied and modeld).


1. What I see as good: Using a white noise model as it is more conservative.

2. What I see as bad:

-- removing the last two observations of a time series. Calling the final observations in a time series outliers is a dangerous game as it could also represent a "turning point" in the series. Several examples show when this happens (from airline passenger rates by year, to the statistical modeling of the Challenger explosion). In any event, it should make us worry about projecting beyond the observed data.

-- What kind of statistician is this??

from the article, after the final graph.

Quote:
That shows just how mistaken, how foolish, how downright boneheaded it is to say that “There is no scientific basis for saying that warming hasn’t stopped.”



The final graph shows that, statistically, we can't conclude any one year since 1993 have been different than zero. ALL of the confidence intervals include zero.

I will give the benefit of the doubt and assume there is some technical detail that has been left out due to the nature of the audience, but it is hard to make the claim --- based on that model -- that warming is still occurring when the last 12 years of data show no significant increase from zero. There is so much uncertainty that there could be a huge amount of warming or even actual cooling -- the modeling being used is not up to the task to answer the question about what will happen tomorrow.

This is the point that DDJ and I always disagreed on -- my argument: the data aren't sufficient to give reliable enough prediction because it is a nonlinear process. DDJ's argument: we don't have time to wait for sufficient data. That decision isn't something that can be answered by statistics.
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
Actually, I agree with a lot of that. Tamino's style is robust and polemical. It needs to be to have any impact.

SpaceGhost wrote:
1. What I see as good: Using a white noise model as it is more conservative.


Red noise actually. Modelling the short term variations as white noise ignores the autocorrelation and leads to underestimates of the uncertainties.

Quote:
2. What I see as bad:

-- removing the last two observations of a time series. Calling the final observations in a time series outliers is a dangerous game as it could also represent a "turning point" in the series.


Really? If there were no extrinsic reason to do so, then I would agree. However, all the months except the last two are based on 14,000+ weather stations distributed globally. The last two data points are based on 47 weather stations in Antarctica. To include those data would require (a) a great deal of trust in the estimation of uncertainties involved in the projection of global temperatures from a single location, and (b) a great deal of trust in some sort of model of the additional contributions to the uncertainties due to the short term variations in temperature.

Quote:
The final graph shows that, statistically, we can't conclude any one year since 1993 have been different than zero. ALL of the confidence intervals include zero.

I will give the benefit of the doubt and assume there is some technical detail that has been left out due to the nature of the audience, but it is hard to make the claim --- based on that model -- that warming is still occurring when the last 12 years of data show no significant increase from zero. There is so much uncertainty that there could be a huge amount of warming or even actual cooling -- the modeling being used is not up to the task to answer the question about what will happen tomorrow.


Yes absolutely. The question needs to be answered from with the science rather than just the statistics. However Tamino is not framing the argument, he's responding to it. Nevertheless, his polemic is too strong.

I've heard it better framed like this:
a) 1999-2010 does not show a trend which is significantly significantly different from either zero or the long term trend. Examining the noise and autocorrelation of this data, we would not expect it to.
b) 1999-2010 forms a sub interval of a period which does show a significantly significant warming trend.
c) 1999-2010 does not form a sub interval of any period which shows a trend which is statistically significantly less than long term trend.

On the basis of standard frequentist tests, surely we can say that we can make no statistically significant claim that warming has deviated from the long term trend?

The alternative is to look at it from a Bayesian viewpoint. I've had no time to calculate this properly, but I did throw together a graph which gives us some hints:


Note that even with the faulty data, the gradient including the last decade (or the last 12 years, the result is even clearer) is greater than the gradient excluding them. The difference is not statistically significant of course. But it suggests that a Bayesian test could easily assign a higher probability to the hypothesis 'warming has accelerated' than 'warming has slowed', although the result will again depend on how the short term variability is modelled.

Edited to add:

Actually, that result is surprising enough to require explanation. Although the gradient on 2001-present is less than the long term gradient, the whole decade is so much hotter than the previous one that adding the extra data increases the gradient. An alternative way of saying this is that the comparative flatness of the last decade appears to be due to a bump above the trend around 2001, not a dip below the trend around 2010. That's why the interpretation depends on how you model short term variations.
11 
 Thumb up
0.01
 tip
 Hide
  • [+] Dice rolls
Daniel Edwards
United Kingdom
London
Unspecified
flag msg tools
eknauer wrote:
myopia wrote:
Pop quiz!

Will Eric:

1. Respond in a reasoned way to the points made in the above post.



Actually, I think Kevin makes it clear that Curry's response was misinformed.



Fair enough, have a thumb.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
United States
Missouri
flag msg tools
mbmbmbmbmb
Kevin C wrote:

Really? If there were no extrinsic reason to do so, then I would agree. However, all the months except the last two are based on 14,000+ weather stations distributed globally. The last two data points are based on 47 weather stations in Antarctica. To include those data would require (a) a great deal of trust in the estimation of uncertainties involved in the projection of global temperatures from a single location, and (b) a great deal of trust in some sort of model of the additional contributions to the uncertainties due to the short term variations in temperature.


In general, I agree. My concern is more if this were to be used for inference/prediction versus description. As I said, I really think that it is hard to argue against increased warming. I think it becomes more difficult when we are doing prediction. If they were abnormalities, then subsequent data points should correct for it and there won't be a need to remove them.

Quote:

Yes absolutely. The question needs to be answered from with the science rather than just the statistics. However Tamino is not framing the argument, he's responding to it. Nevertheless, his polemic is too strong.


Agreed. But, I found his response to be rather one-sided. For instance, the continued claims about the "uncertainity to be even greater" than what is in the model and then only discussing how warming could be almost .3 degrees. To be even-handed, one also has to discuss that the greater uncertainty means that it could also be negative. To present an entire host of models and then discuss them in a biased fashion only undermines the argument.

Quote:

I've heard it better framed like this:
a) 1999-2010 does not show a trend which is significantly significantly different from either zero or the long term trend. Examining the noise and autocorrelation of this data, we would not expect it to.
b) 1999-2010 forms a sub interval of a period which does show a significantly significant warming trend.
c) 1999-2010 does not form a sub interval of any period which shows a trend which is statistically significantly less than long term trend.


That is a much better presentation. The problem is (a) right? Given a large enough variance, anything will not be significantly different than zero or some arbitrarily large number. In a traditional hypothesis testing setting, we would just say that we cannot reject the null that global warming is not occurring (based solely on the 1999-2010 data). The correct response is probably the Bayesian one you discuss below.

Quote:

On the basis of standard frequentist tests, surely we can say that we can make no statistically significant claim that warming has deviated from the long term trend?


I agree with this. The problem is, as you point out above -- you can frame two nulls, right -- (1) recent trend in global warming is not different than zero, and (2) recent trend in global warming is consistent with long term trend. And, we cannot reject either one. The standard frequentist approach can't help us decide yet.



Quote:

The alternative is to look at it from a Bayesian viewpoint. I've had no time to calculate this properly, but I did throw together a graph which gives us some hints:


Note that even with the faulty data, the gradient including the last decade (or the last 12 years, the result is even clearer) is greater than the gradient excluding them. The difference is not statistically significant of course. But it suggests that a Bayesian test could easily assign a higher probability to the hypothesis 'warming has accelerated' than 'warming has slowed', although the result will again depend on how the short term variability is modelled.

Edited to add:

Actually, that result is surprising enough to require explanation. Although the gradient on 2001-present is less than the long term gradient, the whole decade is so much hotter than the previous one that adding the extra data increases the gradient. An alternative way of saying this is that the comparative flatness of the last decade appears to be due to a bump above the trend around 2001, not a dip below the trend around 2010. That's why the interpretation depends on how you model short term variations.


I think this is a much better approach and then the only questions become technical ones. The big one being is how sensitive are the results to the prior? How much data does it take to move away from the prior and what magnitude, for how many years would it need to be. Assuming natural variation levels, can a simulation indicate the timeframe needed for a reasonable negative effect size to change the conclusion? (EDIT: How much of the Bayesian model is predicated on a linear relationship being fitted -- I assume that I can get more abrupt changes if I would use a spline function -- say making the knot at 2001. It seems that the linear model works in conjunction with the Bayesian prior to make it harder to change the conclusion).

In any event, I think it is a step in a better direction -- I also think that it is more fair approach that is stripped of all the rhetoric that only serves to undermine the position.
4 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
United States
Missouri
flag msg tools
mbmbmbmbmb
Kevin -

Is there a place to easily, without much effort, grab the data you showed in the Bayesian analysis?

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
SpaceGhost wrote:
Kevin C wrote:

Really? If there were no extrinsic reason to do so, then I would agree. However, all the months except the last two are based on 14,000+ weather stations distributed globally. The last two data points are based on 47 weather stations in Antarctica. To include those data would require (a) a great deal of trust in the estimation of uncertainties involved in the projection of global temperatures from a single location, and (b) a great deal of trust in some sort of model of the additional contributions to the uncertainties due to the short term variations in temperature.


In general, I agree. My concern is more if this were to be used for inference/prediction versus description. As I said, I really think that it is hard to argue against increased warming. I think it becomes more difficult when we are doing prediction. If they were abnormalities, then subsequent data points should correct for it and there won't be a need to remove them.

Quote:

Yes absolutely. The question needs to be answered from with the science rather than just the statistics. However Tamino is not framing the argument, he's responding to it. Nevertheless, his polemic is too strong.


Agreed. But, I found his response to be rather one-sided. For instance, the continued claims about the "uncertainity to be even greater" than what is in the model and then only discussing how warming could be almost .3 degrees. To be even-handed, one also has to discuss that the greater uncertainty means that it could also be negative. To present an entire host of models and then discuss them in a biased fashion only undermines the argument.

Quote:

I've heard it better framed like this:
a) 1999-2010 does not show a trend which is significantly significantly different from either zero or the long term trend. Examining the noise and autocorrelation of this data, we would not expect it to.
b) 1999-2010 forms a sub interval of a period which does show a significantly significant warming trend.
c) 1999-2010 does not form a sub interval of any period which shows a trend which is statistically significantly less than long term trend.


That is a much better presentation. The problem is (a) right? Given a large enough variance, anything will not be significantly different than zero or some arbitrarily large number. In a traditional hypothesis testing setting, we would just say that we cannot reject the null that global warming is not occurring (based solely on the 1999-2010 data). The correct response is probably the Bayesian one you discuss below.

Quote:

On the basis of standard frequentist tests, surely we can say that we can make no statistically significant claim that warming has deviated from the long term trend?


I agree with this. The problem is, as you point out above -- you can frame two nulls, right -- (1) recent trend in global warming is not different than zero, and (2) recent trend in global warming is consistent with long term trend. And, we cannot reject either one. The standard frequentist approach can't help us decide yet.



Quote:

The alternative is to look at it from a Bayesian viewpoint. I've had no time to calculate this properly, but I did throw together a graph which gives us some hints:


Note that even with the faulty data, the gradient including the last decade (or the last 12 years, the result is even clearer) is greater than the gradient excluding them. The difference is not statistically significant of course. But it suggests that a Bayesian test could easily assign a higher probability to the hypothesis 'warming has accelerated' than 'warming has slowed', although the result will again depend on how the short term variability is modelled.

Edited to add:

Actually, that result is surprising enough to require explanation. Although the gradient on 2001-present is less than the long term gradient, the whole decade is so much hotter than the previous one that adding the extra data increases the gradient. An alternative way of saying this is that the comparative flatness of the last decade appears to be due to a bump above the trend around 2001, not a dip below the trend around 2010. That's why the interpretation depends on how you model short term variations.


I think this is a much better approach and then the only questions become technical ones. The big one being is how sensitive are the results to the prior? How much data does it take to move away from the prior and what magnitude, for how many years would it need to be.



Does this answer your question?

http://www.cato.org/pub_display.php?pub_id=13827&utm_source=...

Quote:
Assume that, beginning in 2013, surface temperatures rise at the rate they did from the mid-1970s through the late 1990s. How much time would have to elapse before a statistically significant warming trend was established, post-1995?

Significance would finally be achieved in 2021. That would be after a 24-year stretch with no significant net warming.

So, in summary, BEST really does not tell us much that is new. It, too, contains the "pause," but the big 1998 peak is missed because the study was only for land areas, while most of the 1998 warming was at the surface of the ocean due to the massive El Niño. Satelllite data, not used in the BEST science, tells us that the East Anglia climate record is probably most reliable for global estimates during the "pause." La Niña conditions mean that 2011 and 2012 are not likely to be particularly warm. Even if surface temperatures begin to rise in 2013 at the rate that they rose from the mid 1970s to the late 1990s, a significant warming trend since 1995 will not be established until 2021.

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
SpaceGhost wrote:
Kevin -
Is there a place to easily, without much effort, grab the data you showed in the Bayesian analysis?

Here's the BEST data: http://berkeleyearth.org/downloads/analysis-data.zip

For a whole load more climate related data, go here: http://chartsgraphs.wordpress.com/climate-data-links/

Note that the BEST data is a land-only index, therefore not directly comparable to the global temperature series more frequently plotted. You can get those for each of the other surface temperature records, but you have to dig a bit. I can help if you are interested.

In fact, now that we have 4 different versions of the land-surface temperature record, it becomes obvious that one of them is an outlier:

The CRU dataset is the only one for which the deviation from the long term trend approaches statistical significance, and then only if you pick the range 1998-2008 - starting on the strongest recorded El-Nino year and ending on a strongish La-Nina year.

That's very interesting. The question is, why? The common explanation has been the lack of arctic coverage in the CRU method, but that doesn't work for the land-only dataset. I'm running some comparisons with my own temperature code to try and clarify why CRU is different.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
eknauer wrote:
Does this answer your question?

http://www.cato.org/pub_display.php?pub_id=13827&utm_source=...

Quote:
Assume that, beginning in 2013, surface temperatures rise at the rate they did from the mid-1970s through the late 1990s. How much time would have to elapse before a statistically significant warming trend was established, post-1995?

Significance would finally be achieved in 2021. That would be after a 24-year stretch with no significant net warming.

So, in summary, BEST really does not tell us much that is new. It, too, contains the "pause," but the big 1998 peak is missed because the study was only for land areas, while most of the 1998 warming was at the surface of the ocean due to the massive El Niño. Satelllite data, not used in the BEST science, tells us that the East Anglia climate record is probably most reliable for global estimates during the "pause." La Niña conditions mean that 2011 and 2012 are not likely to be particularly warm. Even if surface temperatures begin to rise in 2013 at the rate that they rose from the mid 1970s to the late 1990s, a significant warming trend since 1995 will not be established until 2021.



No. Michaels presents no statistical analysis, just a set of assertions. And those assertions are not supported by the data. (At this point I suggested that Micheals knows his assertions were not supported by the data. In fact I do not know if Micheals has or has not done the relevant statistical tests, so I have deleted that assertion. I have done the tests.)

Here are some of the mistakes. The satellite data only shows a pause on the period 1998-2008 if you don't correct for the stronger influence of El Nino on the satellite record, and even that does not approach significance (you need about 17 years of data when using the satellite record, because it is noisier). And even without a correction dealing with the El-Nino effect, the version from the climate skeptic group at UAH shows a stronger trend when you include the recent data:


There are also 5 versions of the satellite record, not two, however only two currently provide monthly data. That is likely to change shortly. The three non-monthly versions all show better consistency with BEST/GISS/NOAA than with HADCRU.

Here's a suggestion: Constantly reposting material which is obviously wrong wastes our time and doesn't do you any credit. To avoid this, you need to be able to distinguish between material which may be correct, and material which is obviously wrong. That's going to require some study and some data analysis, not just googling for articles from climate skeptics.
5 
 Thumb up
1.00
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
Kevin C wrote:
eknauer wrote:
Does this answer your question?

http://www.cato.org/pub_display.php?pub_id=13827&utm_source=...

Quote:
Assume that, beginning in 2013, surface temperatures rise at the rate they did from the mid-1970s through the late 1990s. How much time would have to elapse before a statistically significant warming trend was established, post-1995?

Significance would finally be achieved in 2021. That would be after a 24-year stretch with no significant net warming.

So, in summary, BEST really does not tell us much that is new. It, too, contains the "pause," but the big 1998 peak is missed because the study was only for land areas, while most of the 1998 warming was at the surface of the ocean due to the massive El Niño. Satelllite data, not used in the BEST science, tells us that the East Anglia climate record is probably most reliable for global estimates during the "pause." La Niña conditions mean that 2011 and 2012 are not likely to be particularly warm. Even if surface temperatures begin to rise in 2013 at the rate that they rose from the mid 1970s to the late 1990s, a significant warming trend since 1995 will not be established until 2021.



No. Michaels presents no statistical analysis, just a set of assertions. And those assertions are not supported by the data. i.e. He's lying...

...Here's a suggestion: Constantly reposting material which is obviously wrong wastes our time and doesn't do you any credit. To avoid this, you need to be able to distinguish between material which may be correct, and material which is obviously wrong. That's going to require some study and some data analysis, not just googling for articles from climate skeptics.



I suggest responding to the argument based on the merits and not calling someone a liar because you disagree with them. Do you have proof he is willfully misleading as opposed to being mistaken? If not, then drop the lying accusations. Here's Michael's response to your rebuttal that he is "obviously wrong" (gotta love quick email responses).

Quote:
I am not "adjusting" any data. I'm dealing with the temperature record as it is and merely pointing out how long we are likely to go without a significant trend (at least 21 years). You can fit just about anything to the temperature record--in fact did one of the first publications adjusting temperatures for El Nino, sunspots, and stratospheric dust from 1998-2007 and concluded that those had indeed conspired to stop the warming trend. So it is not like I am not aware of this! That's not relevant to my discussion. If I were comparing MODELED temperatures to observed ones it would be, but I specifically stayed away from that tar sand--I have a paper coming out next year in Journal of Geophysics which does that and concludes that the models are right on the edge (but not quite over) of systematic failure.

If nothing else, this demonstrates that Michaels is not deliberately writing something false that he knows to be true. I appreciate your responses but if you continue with the lying accusations, I don't see any reason to trust your contributions.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
eknauer wrote:
If nothing else, this demonstrates that Michaels is not deliberately writing something false that he knows to be true. I appreciate your responses but if you continue with the lying accusations, I don't see any reason to trust your contributions.


I have edited my post. The bold text describes what I have deleted and why.

I maintain my assertion that Micheals piece contains multiple errors, some of which I outlined above. Interestingly his response introduces a new one: The paper on the 1998-2008 (not 7) trend did not conclude that the trend had stopped (although that misreading is possible if you read only the abstract, which is ambiguously worded).
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
Here's a very interesting graph based on comparison of the Hadley(CRU) and NASA(GHCN) datasets:

Using my own land temperature code, I ran identical calculations on both the GHCN and CRU data. For each dataset I used two different methods:

1. Using an 'equal angle' 5 degree grid to cover the surface of the earth. This is the approach Hadley use, and while simple is probably wrong, because it means that high latitude stations cover a much smaller and elongated region compared to equatorial stations.

2. Using an 'equal area' grid - 5x5 deg at the equator, with the longitudinal width increasing towards the pole. This is the method used by NASA, and gives stations at different latitudes a more equal contribution.

I can see no argument to support the the equal angle approach.

Here's the results:


The dotted lines are the 'wrong' equal angle method, the solid lines the equal area method.

The red lines are the GHCN data used by NASA, the blue the CRU data used by Hadley.

The straight line is just a visual aid (it is the approximate trend since 1975).

Note that agreement is good to 2000. However, using the wrong grid leads to a suppression of temperatures since 2000. Hence, I think there is good reason to consider the Hadley data since 2000 to be suspect.

The difference between the source data is more subtle. The CRU data gives warmer results 2001-2005 only, and this bump at the beginning of the decade is why CRUTEM3 shows a pause in warming. I have no idea which dataset is more realistic at this point, although it has to be noted that the GHCN data is larger, and the BEST dataset larger still.
2 
 Thumb up
1.00
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
Kevin C wrote:
eknauer wrote:
If nothing else, this demonstrates that Michaels is not deliberately writing something false that he knows to be true. I appreciate your responses but if you continue with the lying accusations, I don't see any reason to trust your contributions.


I have edited my post. The bold text describes what I have deleted and why.

I maintain my assertion that Micheals piece contains multiple errors, some of which I outlined above. Interestingly his response introduces a new one: The paper on the 1998-2008 (not 7) trend did not conclude that the trend had stopped (although that misreading is possible if you read only the abstract, which is ambiguously worded).


Thanks, I appreciate the edit and last response.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Kevin C
United Kingdom
York
flag msg tools
mbmbmbmbmb
eknauer wrote:
Thanks, I appreciate the edit and last response.


The thanks are also appreciated. When you've seen the same argument made for the n'th time, it is easy to give a sharp response, but your comments have been courteous and mine have become a bit acerbic. Sorry.

I intend to back up the Hadley work with code (the data is freely available) and will do so when it's done - but I have to hold on to it for now, as it has overlap with a programming assignment I'm setting. I'll start a thread in April 2012 for that purpose.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Eric Knauer
United States
Heathrow
Florida
flag msg tools
mbmbmbmbmb
Kevin C wrote:
eknauer wrote:
Thanks, I appreciate the edit and last response.


The thanks are also appreciated. When you've seen the same argument made for the n'th time, it is easy to give a sharp response, but your comments have been courteous and mine have become a bit acerbic. Sorry.

I intend to back up the Hadley work with code (the data is freely available) and will do so when it's done - but I have to hold on to it for now, as it has overlap with a programming assignment I'm setting. I'll start a thread in April 2012 for that purpose.


Sounds good. Maybe I'll get Patrick to participate sometime. I told him that Jerry Taylor from Cato has been known to post in these forums from time to time although mostly concerning his wargames. I always wondered if the Cato folk even know about Jerry's wargaming geek side.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Prev «  1 , 2  | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.