Guest post by Steve Goddard
Archimedes had his eureka moment while sitting in the bathtub. Newton made a great discovery sitting under an apple tree. Szilárd discovered nuclear fission while sitting at a red light.
There was a time when observation was considered an important part of science. Climate science has gone the opposite direction, with key players rejecting observation when reality disagrees with computer models and statistics. Well known examples include making the MWP disappear, and claiming that temperatures continue to rise according to IPCC projections – in spite of all evidence to the contrary.
Here is a simple exercise to demonstrate how absurd this has become. Suppose you are in a geography class and are asked to measure the height of one of the hills in the Appalachian Plateau Cross Section below.
Image from Dr. Robert Whisonant, Department of Geology, Radford University
How would you go about doing it? You would visually identify the lowest point in the adjacent valley, the highest point on the hill, and subtract the difference. Dividing that by the horizontal distance between those two points would give you the average slope. However, some in the climate science community would argue that is “cherry picking” the data.
They might argue that the average slope across the plateau is zero, therefore there are no hills.
Or they might argue that the average slope across the entire graph is negative, so the cross section represents only a downwards slope. Both interpretations are ridiculous. One could just as easily say that there are no mountains on earth, because the average slope of the earth’s surface is flat.
Now lets apply the same logic to the graph of Northern Hemisphere snow cover.
It is abundantly clear that there are “peaks” on the left and right side of the graph, and that there is a “valley” in the middle. It is abundantly clear that there is a “hill” from 1989-2010. Can we infer that snow cover will continue to increase? Of course not. But it is ridiculous to claim that snow extent has not risen since 1989, based on the logic that the linear trend from 1967-2010 is neutral. It is an abuse of statistics, defies the scientific method, and is a perversion of what science is supposed to be.
Tamino objects to the graph below because it has “less than 90% confidence” using his self-concocted “cherry picking” analysis.
So what is wrong with his analysis? Firstly, 85% would be a pretty good number for betting. A good gambler would bet on 55%. Secondly, the confidence number is used for predicting future trends. There is 100% confidence that the trend from 1989-2010 is upwards. He is simply attempting to obfuscate the obvious fact that the climate models were wrong.
Science is for everyone, not just the elite who collect government grant money. I’m tired of my children’s science education being controlled by people with a political agenda.
“The trend from 1989-2010 is clearly upward at 100% confidence.”
I’m sorry. That’s just not how it works.
“Here is an exercise for Robert et. al”
Who are my et al? Have I acquired a scientific posse without realizing it?
“Show me statistical significance in the geologic record between CO2 and temperature.”
I am not the whiz with computers that some people here are, but I’d be happy to take a look.
This is reminiscent of the controversy that surrounded the claim that snowpack in the Cascades shrank by 50 percent in the last half century. The statistic was used by government and advocacy groups to push their climate change agenda in Washington State. The only problem is that the statistic was dead wrong. The unfolding scandal revealed extensive corruption of the science by one researcher, Philip Mote, at the University of Washington. The Seattle Times published a good summary of the story when it first broke a few years back:
http://seattletimes.nwsource.com/html/localnews/2003618979_warming15m.html
The same story was revisited by the Air Vent in July of 2009 and Jeff Id carefully dissected the methods used by Mote et al to reach their fraudulent conclusions. The article is well worth reading given the many similarities between Mote’s dubious methods and the methods purportedly being used to discount NH snow extent:
http://noconsensus.wordpress.com/2009/07/21/snowmen/
It’s also interesting to note that just a few months ago, the odious NPR environmental reporter, Richard Harris, gave an extremely distorted picture of the Climategate scandal on the NPR program “All Things Considered”. That by itself is not terribly surprising. However, what shocked me most during the segment was how Harris completely misrepresented John Christy’s contention that his research paper on Sierra Nevada snowpack melt had been suppressed for political reasons. Harris casually concludes that Christy’s concerns were simply a matter of paranoia stemming from Christy being personally attacked in the CRU emails.
Please read the transcript (below) and carefully note who actually reviewed and rejected Christy’s paper on snowpack melt. Then reread the article I posted above on the Cascade snowpack fraud. Finally, revisit Christy’s contention that his paper was intentionally suppressed for political reasons. See if you reach the same conclusion that Harris does and report back.
TRANSCRIPT:
That said, many of the complaints about the journal review process come from people who think mainstream science is overstating climate hazards. John Christy is at the University of Alabama at Huntsville. He accepts global warming is happening, but he says there’s a lot of uncertainty about its causes and impacts. And he says he has trouble getting some of his results published.
Professor JOHN CHRISTY (Director of the Earth, System Science Center, University of Alabama, Huntsville): I’ve done a pretty thorough study of snowfall in the Sierra Nevada mountains of California. And the Southern Sierra show no downward trend in snowfall.
HARRIS: That’s important because snowfall is forecast to decline due to global warming. And that would seriously affect California’s water supply. Christy says he’s tried three times to get his paper published. So far it’s been rejected, and he suspects it’s because scientists are trying to stifle his message.
Prof. CHRISTY: Everyone from the secretary of energy who has talked about the snowfall in the Sierra going away will not find any comfort in the fact that the trends in snowfall are essentially zero for the last hundred years.
HARRIS: So is it being suppressed? Philip Mote at Oregon State University was one of the scientists who reviewed the paper. He said the science in the paper was fine.
Dr. PHILIP MOTE (Oregon Climate Change Research Institute, Oregon State University): To my knowledge, there’s no suppression going on. It’s simply that it’s not news.
HARRIS: Mote himself published a paper four years ago showing that snowfall in the Southern Sierra hasn’t diminished. In fact, he says there are about 10 papers on the subject, certainly not identical to Christy’s but still reaching that same broad conclusion.
Dr. MOTE: It’s not controversial because it’s already well known.
HARRIS: Still, it’s easy to see why Christy suspects deeper motivations. The stolen emails contain sharp personal attacks against him. He says the politics inside climate science are making life harder for him, not just in publishing papers but in getting money to do research. Mote says dissent is important in science. He doesn’t agree with everything Christy says, but he says he should have a voice.
—-
Full transcript can be found here:
http://www.npr.org/templates/transcript/transcript.php?storyId=120846593
B.D.
You are the voice of a true scientist! Someone who reads, thinks and observes carefully, and trusts his own eyes.
R Gates
“So if the trend toward more snow covrer (DURING WINTER) is accurate”
The trend is moving south. It’s called extent.
That means it’s colder further south, where it was not colder before.
Moisture without cold, is called rain.
When the extent of snow cover increases, that means it’s colder.
Re: Robert (14:20:54) :
I don’t understand why you think that the 95% confidence limit or the 0.05 significance level is there somehow by definition in statistics. Confidence limits are arbitrarily set. They are subjective not objective. There is nothing in the definition of statistical significance which demands that a certain significance level be used. It just so happens that people have come to use the 0.05 level of significance for many studies. That doesn’t mean that people are right to use 0.05 just that a lot of people do. Something can be statistically significant at the 85% confidence interval. It may not be significant at the 95% CI but it doesn’t change the fact that by the very natural of statistics that it is statistically significant at the 85% confidence interval. Saying something is statistically significant is not a valid statement unless it is paired with the levle at which that significance was tested.
To claim a 95% confidence interval is correct because of definition is ludicrous because no such definition exists. If you would like to present an argument for why 95% is the holy grain of significance levels other than that because when it comes to plucking significance levels out of thin air most people tend to pick 95% out of tradition then I would like to hear it. Why 95%? Why not 96%? Why not 94%? What is so special about 95% which, regardless of how many people arbitrarily select it, doesn’t make its selection arbitrary?
“Show me statistical significance in the geologic record between CO2 and temperature.”
Here you go. Somebody else has done the work:
http://4.bp.blogspot.com/_KfE5s-4q1s4/S0PxgIwS7BI/AAAAAAAAAD0/ydkuhHeRAZg/s1600-h/co2vstemp_glacial.jpg
You will note that R^2= 0.79, and n = 799. That’s about as significant as they come. Cf this handy chart:
http://www.mega.nu/ampp/rummel/uc.tab9.2.gif
Description of the analysis: http://moregrumbinescience.blogspot.com/2010/01/co2-and-temperature-for-800000-years.html
BerneiL
Using bar charts makes it difficult to determine trends.
Scatter charts are better for that putpose (snow cover, temperature or whatever on the Y axis, time on the X axis).
From a practical point of view, I still prefer line charts, as I am used to intrepreting these.
But if you are picky, then use scatter plots.
I think this is funny. Does anyone remember why statistical analysis and confidence levels and so on are so important? I shall answer my own question.
Human being percieve data differently than computers do. The value of computers is that a) they can manipulate data FASTER than human beings can and b) they can manipulate MORE data than human beings can. The problem this creates is that we no longer percieve the data directly, we percieve it through the limitations of the computer analysis that manipulates it for us. Anyone who has worked for example, in a custom metal fabrication facility, knows that drawings come in from engineers who used Autocad that could have been far simpler to fabricate other ways. I know a professor of architecture who sends his students to a particular building here in town and asks them to use Autocad to draw the stairwell. They can’t, which is the purpose of making them try and do it. Computer tools allow analysis that human beings could never accomplish with calculators or pencils and paper. But it also means we need mechanisms to validate that what we see in the computer analysis output is what we thing it is. Hence statistical analysis and confidence levels to understand if the graph we are looking at, which may be several generations of analysis from the original data, says what our human perceptions believe it says. There’s too much data for us to look at several million measurements, a graph the computer built, and go… yeah, that looks about right.
The point of that ramble, is that human perception and analysis is still superior to that of a computer program, provided that the data being observed and considered is small enough for us to “fit it in our heads”. In Steve Goddard’s example of the hills, this is exactly the case. Take ten thousand altitude measurements across those hills, and a computer program will in fact tell you that the average is flat, that there is no trend, that the variability is within error, and it will be correct. That’s how computers see data and that’s why it is so easy to take big chunks of data and get answers that are technicaly accurate and totaly meaningless.
Anyone can look at the graph of the hills and answer the question. Yes they need to say compared to what? the valley? see level? where I’m standing? But it is obvious that saying there are no hills at all, its flat, could be accomplished ONLY by computer analysis, not by a human being looking at it.
I don’t need a computer to analyze 40 years of snow extent data. I might have some questions as to how it was measured and so on, but having been satisfied that the data collected is valid…. looks to me like snow extent is highly variable, that there was a declining trend from the beginning of the graph that was accompanied by decreased variability, followed by an increase in snow extent at the end of the graph that seems to be accompanied by increasing variability.
We’ve gotten so used to looking for the errors that poor math or bad data in conmputer models can produce, that we forgot the purpose was to ensure we were looking at what we thought we were looking at, and if the data set is something we can percieve in its entirety without a computer, then there is no need for the statistical analysis. We can actually just look at it.
Robert,
Nice try. That wasn’t the question I asked. The data you provided demonstrated that CO2 concentration follows temperature in the narrow range of the glacial record over the last few hundred thousand years. During that time CO2 only varied in a very narrow range.
I pointed you to a graph of the geologic record going back 600 million years plus where CO2 varied by 1000-2000%, which shows very little if any correlation between CO2 and temperature. Please try again.
http://i224.photobucket.com/albums/dd137/gorebot/Geological_Timescale_op_927x695.jpg
Here is another one.
http://ff.org/centers/csspp/library/co2weekly/2005-08-18/dioxide_files/image002.gif
Over geologic time, there is little if any correlation between CO2 concentration and temperature.
‘There is 100% confidence that the trend from 1989-2010 is upwards.’
I’m intrigued by this statement Steve. The main problem I have with this graph is the lack of error bounds. In my experience I have seen many wonderful ground breaking trends which magically disappear once I add the error bound along with my hopes of a tier 1 journal paper.
Even with empirical data there is still the question of how well it represents reality. This is why I disagree with another statement you made about statistics only being for predicting future trends. Carl Popper says that we use statistics to make up for a lack of knowledge and I’m sorry to say that I like his interpretation better.
If you make three observations under the same conditions and get different results then there is something you don’t know. Now you want to change the conditions and see if you can measure a change in the result. The problem is that the result changes even when you don’t change the conditions because of what you don’t know. So how can you say anything about changes in results due to changing conditions when results change without conditions? Well we all know the answer to that one and it is statistics. You compare the changes that occur under constant conditions (what you don’t know) with the changes that occur under changed conditions (the sum of what you do and don’t know) and compare them to see if you know enough to make inferences.
I don’t think this process involves predicting future trends, its simply analysing past events for significant patterns.
Robert re your evidence for AGW.
The missing gap is how the other, far more important players in the climate interact with the effect of CO2.
Some of these include water vapour fluctiations, cloud cover, earth – atmosphere -ocean interchanges and cosmic rays (currently being tested in the Cloud experiment at CERN). There are probably others I have omitted.
The problem I have with AGW is that I have taken NCDC temperature data from 1880, removed the long term linear trend and the 65 year zig zag cyclic component. I am left with short term fluctiations cause by el Ninos, volcanos and higher harmonics of the 65 year cycle.
These short term fluctiations have no trend (r squared value very close to zero).
There is no visable evidence of any CO2 effect in these residuals.
(You will see that I have completely ignored the increasing evidence that the basic temperatrure data may have been biased upwards systematically, which would have artifically raised the long term trend.)
On top of that is the evidence, growing daily, that the supporting arguements for glacier loss, floating ice loss, snow loss, drought, fire flood, storms, rate of increase in sea levels, polar bears etc are all either exagerated or completely false. That suggests that the scientists at the heart of the AGW scare are pursuing an agenda rather than scientific research.
I do not consider myself a denier or a skeptic. My position is that I do not see sufficient evidence that AGW is real. I am an accountant by profession with both econimic and statistical training. I realise the devastation that the proposed massive reductions in using fossel fuels woud cause.
That deveatation would be far, far more severe than the goverments around the world are prepared to admit, or perhaps even realise. Perhaps that is because so few politicans have no experience of the difficulties of actually creating wealth that all of us rely on to survive.
Making the huge sacrifice of abandoning carbon would be extremely difficult, with misery for all and death for many. To make such monumental change without good cause would be completely wicked.
I’d just like to state for the record that the Wikidpedia picture of Sir Issac Newton looks remarkably like a guy who works in my community theatre.
R. Gates (16:29:54) :
“How that discredits AGW in any way is beyond me…”
==============
How do you define AGW ? (Show your work).
Robert (12:09:01) :
Excellent. Then we can ignore the IPCC AR4 almost completely, as it almost never reaches 95% (“extremely likely, by their assessment) confidence levels.
Thanks.
Can I interest you in a game of chance, by the way, only $10 a point…..
Steve Goddard (16:24:35) :
Most school kids would understand the concept of measuring the height of a hill, person, tree, building, television, etc. How about you?
It becomes a bit more tricky if the person, etc is standing on a steep slope. Perhaps there are several persons. One could define the highest person as the one being being able to see farther. A valid analysis would have been simply to plot your curve on the same graph as the models, lining them up on some point in the past, e.g. when the model was run or the last data point used as input to the model. Then there would be no discussion, anybody could just by eye get a feeling for how well the models were doing. Try that.
This is reminiscent of the controversy that surrounded the claim that snowpack in the Cascades shrank by 50 percent in the last half century. The statistic was used by government and advocacy groups to push their climate change agenda in Washington State. The only problem is that the statistic was wrong. The unfolding scandal revealed extensive corruption of the science by one researcher, Philip Mote, at the University of Washington. The Seattle Times published a good summary of the story when it first broke a few years back:
http://tinyurl.com/259rsh
The same story was revisited by the Air Vent in July of 2009 and Jeff Id carefully dissected the methods used by Mote et al to reach their fraudulent conclusions. The Air Vent discussion is well worth reading given the many similarities between Mote’s dubious methods and the methods purportedly being used to discount NH snow extent:
http://tinyurl.com/ydp58zf
It’s also interesting to note that just a few months ago, the odious NPR environmental reporter, Richard Harris, gave an extremely distorted picture of the Climategate scandal on All Things Considered. That by itself is not terribly surprising. However, what shocked me most during the segment was how Harris completely misrepresented John Christy’s contention that his research paper on Sierra Nevada snowpack melt had been suppressed for political reasons. Harris casually implies that Christy is a conspiracy theorist and that his paranoia stems from being personally attacked in the CRU emails.
Let’s do an experiment. Since Richard Harris is apparently an incompetent journalist, let’s revisit Christy’s contention that his paper was intentionally suppressed for political reasons. First, please read the article linked to above about the fraudulent claims made by Philip Mote regarding an increasing trend in snowpack melt in the Cascades. Next, please read the NPR transcript below and carefully note who actually reviewed and rejected Christy’s paper on the lack of a downward trend in Sierra Nevada snowfall. Finally, see if you are able to reach the same conclusion as Richard Harris and report back.
_____
TRANSCRIPT:
That said, many of the complaints about the journal review process come from people who think mainstream science is overstating climate hazards. John Christy is at the University of Alabama at Huntsville. He accepts global warming is happening, but he says there’s a lot of uncertainty about its causes and impacts. And he says he has trouble getting some of his results published.
Professor JOHN CHRISTY (Director of the Earth, System Science Center, University of Alabama, Huntsville): I’ve done a pretty thorough study of snowfall in the Sierra Nevada mountains of California. And the Southern Sierra show no downward trend in snowfall.
HARRIS: That’s important because snowfall is forecast to decline due to global warming. And that would seriously affect California’s water supply. Christy says he’s tried three times to get his paper published. So far it’s been rejected, and he suspects it’s because scientists are trying to stifle his message.
Prof. CHRISTY: Everyone from the secretary of energy who has talked about the snowfall in the Sierra going away will not find any comfort in the fact that the trends in snowfall are essentially zero for the last hundred years.
HARRIS: So is it being suppressed? Philip Mote at Oregon State University was one of the scientists who reviewed the paper. He said the science in the paper was fine.
Dr. PHILIP MOTE (Oregon Climate Change Research Institute, Oregon State University): To my knowledge, there’s no suppression going on. It’s simply that it’s not news.
HARRIS: Mote himself published a paper four years ago showing that snowfall in the Southern Sierra hasn’t diminished. In fact, he says there are about 10 papers on the subject, certainly not identical to Christy’s but still reaching that same broad conclusion.
Dr. MOTE: It’s not controversial because it’s already well known.
HARRIS: Still, it’s easy to see why Christy suspects deeper motivations. The stolen emails contain sharp personal attacks against him. He says the politics inside climate science are making life harder for him, not just in publishing papers but in getting money to do research. Mote says dissent is important in science. He doesn’t agree with everything Christy says, but he says he should have a voice.
_____
Full transcript can be found here:
http://tinyurl.com/ydk64bu
“But it is believed that the process is making the seas much more acidic which is damaging the delicate shells of organisms that are critical to the marine food chain.”
Drivel. Unsubstantiated propaganda.
Steve Goddard (17:55:23) :
Over geologic time, there is little if any correlation between CO2 concentration and temperature.
Methinks I showed you there was, when you asked me to ‘prove it’… Have you already forgotten?
Robert, thank you so much.
Since you brought all of this up about r2 I have learned a big lesson. R2 is never an indication of confidence. It is a description of slope when using a linear trend. A flat trend line ALWAYS has an R2 of 0. A near vertical trend line ALWAYS has an R2 near 1. Bingo. It doesn’t matter how many of where the data points are around the trend line when speaking of a time-series. Robert, keep up the good work.
Everone, try it in Excel. Your view of R2 will never be the same again.
Steve Goddard:
Actually Mark Twain attributed the quote to British PM Benjamin Disraeli, except that Disraeli (according to his biographers) never made such a statement.
That’s why you should check your sources.
Statistics is not a concentrated form of lying, but liars can make fools of us all by bending statistical analyses to a particular conclusion.
Statistics is a mathematical formalism that asks questions about data and trends which could arise by chance.
So for example, a time series which exhibits long term persistence will easily produce short term trends which could be the result of a stochastic process.
You have produced a model showing an apparent linear trend in a time series. Is the series significant? There are a battery of tests created by statisticians which give a likelihood as to whether your claimed trend could have a risen by chance.
Berating the rest of us for grasping that point because it eludes you is neither skeptical nor scientific.
Leif Svalgaard (12:34:51) :
“I’m tired of my children’s science education being controlled by people with a political agenda.”
Move to Kansas…
and since I happen to find evolutionary theory convincing…what say you now?
wayne (19:14:27) :
A near vertical trend line ALWAYS has an R2 near 1. Bingo. It doesn’t matter how many of where the data points are around the trend line when speaking of a time-series.
None of this is correct.
Steve Goddard said: (17:07:57) :
“R. Gates,
Do you understand the difference between snow depth and extent? Extent increases when it snows in Florida. Snow in Florida is not due to heat – it is due to cold.”
Again with insults? Why? You know that cold is only half the equation to produce snow. All the moisture that fell on the south in the form of snow came from subtopical waters– either the Gulf or the Pacific or both. Generally speaking, COLD=DRY in climate terms. This is climatology 101.
But to answer your question: You know full well why there was enough cold in Florida for it to snow– a very negative AO. And if you’ve done lick of homework, you also know that negative AO’s have their ROOTs not in the arctic, but in the the tropics where warm air is forced up through the troposphere and carried northward to decend on the arctic, forcing the cold arctic air south, and creating high pressure over arctic regions in the process as well. I’ve displayed previously the charts showing the high pressure we had over Greenland this winter with the concurrent higher temps, when it was warmer in Greenland than in much of the southern states.
Again, deep snow, more snow, gtreater extent, however you want to measure it has it roots in HEATING of the oceans. You are somehow trying to equate a greater snow cover with a colder overall climate…and that’s just plain wrong.
Anarctica is cold and dry
The last glacial period was cold & dry
Cold=dry and Warm=wet
The fact that we have an El Nino year, plus record temps in the troposphere, plus the several negative AO events, (related to warm tropics) is precisely why we had big snows in odd places this year, and so, if your charts are correct, and we’ve had more snow since 1989, then that only proves there is more heat around in the winter to evaporate and transport all that moisture.
The simple (and wrong) approach that cold=more snow, defies the laws of both physics and the history of the climate on earth.
Leif Svalgaard (19:33:23) :
wayne (19:14:27) :
A near vertical trend line ALWAYS has an R2 near 1. Bingo. It doesn’t matter how many of where the data points are around the trend line when speaking of a time-series.
None of this is correct.
I should rephrase this. A trend line that is significant enough has an R2 near 1. Does not have to be vertical. Try in excel the data 2 4 6 8 10 in one column and 1 2 3 4 5 in another and see what R2 is. Time series has nothing to do with anything.