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.
Steve Goddard (08:47:02) :
Seven out of the last ten years have been above 45 million km2, and eight out of the previous eleven were below 45 million km2. That is an upwards trend, ?que no? Speaking of statistics, that is a several sigma event.
Now we are back to Hemispheric data, not N.A. [remember?] that you started with. And your above/below ‘analysis’ is not valid unless you quantify how much above/below. Suppose there were only 1 km^2 above/below. The so difference has to be quantified and compared to the usual noise in the data. There are standard methods of doing that [Tamino is good at it 🙂 ]. Lean about them and try them.
Leif @ur momisugly 8:57:57
Firm, patient, and silent, he ain’t. Noisome.
======================
Leif,
So you agree with Tamino that winter snow extent has not increased over the last twenty years?
Leif Svalgaard (08:57:57) :
You see, you just can’t admit it can you?
You just keep avoiding the questions.
I am not talking Trends I am talking about reality and you just don’t get it, WHY?
Please answer those 3 questions as I put them.
I will answer yours about the 3 odd Years and yes they had less snow.
Your turn, yes or no?
Steve Goddard (09:13:36) :
So you agree with Tamino that winter snow extent has not increased over the last twenty years?
Statistically it hasn’t, because the change is compatible with the natural variation and noise. A single year’s data cannot be said to be ‘statistically significant’. For a series of data points one can ask if the trend in the data is statically significant. That is not the same as whether it is ‘real’. The data are what they are. The question is if one can derive a climatologically significant trend from the series. And Tamino is right, one cannot from the data presented. But that is not your real issue, which was [remember?] whether the models were correct in their predictions and it certainly looks to me that they were not, because the divergence between the two data sets is probably too large. And the downward trend in the models does seem to be statistically significant because the variation in the prediction is rather small [although from one model to the next, but within one model], so even if the observations are consistent with no trend [flat], the models are not. Now, the decrease in summer snow cover is worrisome for you [or should be], being a more sensitive indicator of climate. Glaciations come about when the snow at higher latitudes begin not to melt in the summer.
A C Osborn (09:30:44) :
I am not talking Trends I am talking about reality and you just don’t get it, WHY?
See reply 09:45:07 to Steve
Leif,
The last few years have seen exceptional snowfall in the Northern Hemisphere. People shoveling snow don’t care about scientists playing games with statistics
Changes in summer snow/ice cover which occurred prior to 1989 (as I have written many articles about) is largely due to soot, Hansen/Nazarenko, Zander etc.
Note Tamino’s graph showing that August snow cover is about the same as twenty years ago.
http://tamino.files.wordpress.com/2010/02/nhaug.jpg
Leif Svalgaard (09:46:41) :
See you still can’t do it.
Perhaps if that was your income and I subtracted those amounts from it (instead of added those kinds of amounts), perhaps you might find it “SIGNIFICANT” then, as the people living in the NH are finding it now with Snow blighting their lives.
There is just no point in reading what you have to say as you are evasive in your answers.
Steve Goddard (10:01:23) :
I don’t know why you are still bothering, Leif is not going to agree unless we have at least another 10 years of heavy snow, but I admire you trying.
A C Osborn,
Scientists love math and statistics. They are very important when appropriate, but sometimes people just need to look out the window. ;^)
Steve Goddard (10:01:23) :
Note Tamino’s graph showing that August snow cover is about the same as twenty years ago.
http://tamino.files.wordpress.com/2010/02/nhaug.jpg
but is much less than what is was 19 years ago.
Steve Goddard (10:01:23) :
The last few years have seen exceptional snowfall in the Northern Hemisphere. People shoveling snow don’t care about scientists playing games with statistics
I took one of the models you were talking about. It is here:
http://www.leif.org/research/Snow-Cover-1850-2100-Model.png
According to the model there should be a significant increase in snow cover during 1985-2020, as observed. [my arrow in the box]. This is to show that you can’t draw conclusions on a few years of data. The variability is just too large. That is what statistics can tell you.
If you are only “talking about reality” then it is obvious from your graph that some years it is higher than in some previous years and some years it is lower. So if you say, for example, “it was lower in 2009 than it was in 2008,” in the same way you might say, “it was cloudy yesterday but clear today,” without pretending that it has any meaning beyond idle observations about the weather, that is fine.
On the other had, if you want to say something general about what is happening overall (e.g. “snow cover has been increasing/decreasing”), or suggest that the measurements have any meaning, then you are doing statistics, and you need to do it correctly. Leaving out considerations of trend or statistical significance or the impact of choosing to analyze only one particular subset of the time record does not get you away from doing statistics–it just means that you are doing bad statistics.
trrll (11:37:57) :
does not get you away from doing statistics–it just means that you are doing bad statistics.
Or none at all, just looking out the window.
sometimes people just need to look out the window.
If I could see the whole Northern Hemisphere from my window – and if my window had a “rewind” button – you might have a point here. Otherwise, this is a non sequitur/i>.
From my window in Boston, I’ve hardly seen any snow at all this winter. Washington is up to something like five feet.
Steve Goddard (10:01:23) :
don’t care about scientists playing games with statistics
I took one of the models you were talking about. It is here:
http://www.leif.org/research/Snow-Cover-1850-2100-Overlay.png
and overlaid the snow cover data [the heavy orange graph]. You can see that the model does a reasonable job, or rather that the variations [error bars] are so great that even if it looks good you can’t really make any conclusions. That is what statistics also would tell you.
Leif, I have a question. Why do you call the variation of the curves an “error bar”?
I do not know how the plot was made. I would expect that somebody integrated the area covered by snow. The error on this measurement will be something like maybe 1 kilometer? 10 kilometers? I would not expect it to be larger as satellites are accurate.
It is a variation, maybe we cannot explain it, but it is not an error in the measurement. The peaks and troughs are probably chaotic , but probably also, in the same way that one can have large waves and small waves on average depending on the wind, one can have high snowfalls and low snowfalls depending on the ENSO or whatever alphabet combination. I would not call them an error.
Now the model, which by construction of GCMs purports to predict a curve, does not have errors either. And the errors it does not have are much more important, because they do no error propagation from their parameters. So they throw a curve on the table with wiggles. One is legalized to look at the total wiggle pattern, and the trend in the data and the model but not to talk of the variations as errors.
I agree that there are not enough peaks and troughs to say whether it will be probably rising in the future, in the same way that watching a few high waves does not define that the wind is rising. If they get to be 100 and growing , that would give the relevant statistical error.
Dang, all those news stories the last few years about record snow in Texas, Florida, Rome, Iraq, Saudi Arabia, Buenos Aires, etc. must have been lying. The Rutgers data about near record snow extent in 2008 and 2010 must also be wrong.
We have statisticians who say otherwise. Can I make a suggestion? If you want to pick a fight about trends in snow cover, don’t do it during a period of record high snow extent.
Is winter snowfall going up fro the last 20 years, or is it going down?
http://wattsupwiththat.files.wordpress.com/2010/02/north_american_winter_dec-feb_snow_extent_1989-2010.png
Talk, blah, blah, statistics blah, blah, blah, cherry picking, blah blah blah …….
What advice would Kartman give?
anna v (13:01:33) :
Why do you call the variation of the curves an “error bar”?
Poor choice of words that then takes on a life of its own. What I meant [and mistakenly assumed was obvious] was that with a regression line, you can draw two other lines [slightly curved], one on either side that shows the confidence ‘bands’ for a given level of confidence. You can draw several such sets of curves for different levels, 99%, 95%, 90%, 68%, etc. The width of each band is what I meant by using the shorthand ‘error bar’.
Statistics are not helpful in a situation where the physical processes involved are unknown.
Instead one can legitimately employ informed judgement.
It is clear that global tropospheric temperatures cycle up and down but are subject to a good deal of erratic or chaotic behaviour in the process.
So whatever statistics say it is becoming quite clear that a peak of global temperature was recently passed and we could now be on an accelerating downslope. The current extent of northern hemisphere snow cover is one of a number of real world phenomena pointing in the same direction regardless of statistics.
Steve is quite right to point to the past 20 years of slow but still small increase in snow cover over time at certain parts of the year.
Arguing that a single parameter is not statistically significant on it’s own is fair enough but we actually have several such indicators all pointing the same way and taken as a group I suggest there is statistical significance or if there is not then the statistical methods need adjusting rather than our powers of judgement.
“Lies, damned lies and statistics.”
That is a judgement call and statistics don’t help.
I have been following the conversation between you two, and I have one thing I’d like to add. It is darned difficult to identify a trend (secular increase) in data known to contain cycles. We have this same problem in manufacturing. The idea of having a 30 year dividing line between climate and weather is fine, but if a very significant cycle is longer than 30 years, maybe we should make the dividing line at this cycle length. So, for example, the PDO cycle length looks to be 60 years to me, so why wouldn’t we try to enclose it so we don’t chase our tails running after climate change that is actually cyclic? Make weather 60.
The problem may be that we never find a longest cycle to encompass, but that would be very telling in its own right wouldn’t it?
What advice would Kartman give?
Really???
You want to go there ???
Steve Goddard, I’m glad you brought up the Ordovician Ice Age: It did indeed occur sandwiched between Greenhouse climates. But this glacial episode only lasted a few million years at the most, and was possibly as short as 500,000 years. But it was also very severe. Continental rearrangements can’t be the answer, as they happen slowly, and thus wouldn’t have changed dramatically for 10 million years either side.
It appears that CO2 may be the answer after all:
http://www.eurekalert.org/pub_releases/2009-10/osu-vpp102609.php
Turns out that if you look at detailed isotope records, there was a large amount of volcanism along the proto-Atlantic margin, causing the high CO2 levels you mention, and a corresponding Greenhouse Climate. But these emissions were somewhat held in check by the massive erosion of the uplifting Appalachian mountains which sequestered a lot of CO2. Then, for some reason, the volcanism abruptly stopped, but the weathering continued, causing a massive drawdown of CO2 from the atmosphere. The Hirnantian Ice Age was the result.
So one lesson to take away from this is that you should be more skeptical, rather than accept graphs that suit your idea of how things work.
Stephen Wilde (14:22:05) :
Statistics are not helpful in a situation where the physical processes involved are unknown.
On the contrary, that when we need statistics the most. If we know the physics we don’t need statistics.