Apparently, Roy Spencer believes that warming indicated by data from the USHCN (U.S. Historical Climate Network) is almost entirely false. He likewise distrusts the trend for the U.S. estimated from the CRUTem3 data. More to the point, he seems to think that warming over the U.S. for the last several decades has been negligible. All but a pittance (a mere 0.013 deg.C/decade), he says, isn’t real, it’s just due to “adjustments.”
Speaking of temperature trend for the U.S., here’s his claim in detail:
The linear warming trend during 1973-2012 is greatest in USHCN (+0.245 C/decade), followed by CRUTem3 (+0.198 C/decade), then my ISH population density adjusted temperatures (PDAT) as a distant third (+0.013 C/decade).
The curious thing is, if you look at the satellite data for lower-troposphere temperature from UAH (Spencer’s own data set), the trend since 1979 for the U.S. is 0.22 deg.C/decade:
The UAH data indicates that the troposphere over the USA is warming 17 times as fast as Spencer’s “PDAT” trend. Even if we allow for uncertainty in the trend from UAH data, we have a 95% confidence interval from 0.088 to 0.356 deg.C/decade. So the tropospheric warming rate might be only 6.8 times as fast as Spencer’s “PDAT” trend. Then again, it might be 27 times as fast.
Does Spencer really believe that the troposphere over the U.S. is warming 17 times as fast as the surface? Or does he not trust his own UAH troposphere data set? It’s a mystery.
Open Mind 
After reading Spencer’s silly post I downloaded the UAH data, ran a linear regression for the same period and of course got the same result as you. It’s not such a big deal to do.
I think someone may have made a comment at WUWT about the weirdness of his publishing one set of numbers yet condemning another set of numbers that show the same thing. But the vast majority of commenters there just don’t get it. What a freakshow.
(As an aside, I learned about your blog after I had been pointed to one of Spencer’s posts at WUWT. Realizing Spencer didn’t understand least squares regression I explained to the fellow what was wrong with his reasoning. Another fellow came into the discussion and pointed me to your “Bag of Hammers” post, saying you’d concluded the same thing. And that’s when I became a regular reader here.)
Having had a go at correcting a few inaccuraces in a thread on WUWT back in January, I was set upon by a gang of Watt’s disciples, egged on by Watts himself. It wasn’t a pleasant experience as it was 90% personal abuse and only 10% discussion of the science. I soon realised why so few people stick around to disagree with posts on that site.
The mutual mathturbation indulged in by the scientifically-illiterate on there has to be experienced to be believed. What’s amazing is that you can write any old tosh you like and, as long as you appear to rubbish climate science, you’ll be applauded whatever you say: certainly you’ll never be corrected, however ludicrous your assertions!
Had similar experiences at Curry’s Climate Etc..
It is disturbing that people can engage in that sort of behavior without the slightest trace of self-consciousness..
Great Post.
Spencer’s post and the reaction at WattsUp are unbelievable!
Currently there are 126 comments on Spencer’s post at WattsUp. One poster mentioned that the UAH trend showed an increase of 0.5C over 30 years. One of the so called skeptics replied that the next 30 years of PDO would wipe out this trend, which was quite beside the point.
Another poster pointed out the UAH data doesn’t agree with Spencer’s ISH data set. The moderator berated him for being a troll.
The self deception of the socalled “skeptics” is amazing. Most of the posters there are high fiving Spencer, and ignoring the discrepancy.
I’ve been trying to follow roy’s method in order to understand exactly what this population density adjustment is, but he crucially leaves out the equation that he uses to “remove (this average relationship between temperature trend and population density) from each station”.
The relationship apparently is:
Trend (deg C per decade)= 0.0586x + 0.0128
x = (pop per km^2)^0.2
I suspect that he the simply subtracted the “0.0586x” part from each station (or something very similar), i.e.,
New_station_T = Old_station_T – 0.0586*x*(year – 1992)
Because 1992 is the mid point of the range 1973 to 2011.
This would appear to guarantee that the final adjusted trend will be very close to the left over “0.0128” deg C per decade part of the “average population density relationship”.
Sure enough, it worked out to 0.013 deg C per decade!
Otherwise, I don’t see how he can take so many stations that almost all show significant warming trends at all population densities, even as low as 1 per km^2, and produce a temperature data set with almost no trend.
If this is what he did then it makes no physical sense. First, it appears to imply that a constant population density causes temperatures to steadily increase for decades and that going from zero population to 1 per km2 has the same warming effect as going from 1000 per km2 to 3000 per km2.
How one person can cause an entire km2 of surface area to warm up indefinitely at the rate of almost 0.6 deg C per century doesn’t appear to be an interesting question for dr Roy.
I would like to point out that Spencer has a history of not looking at the physical reality of his claims outside the peer reviewed literature (not that his actual publications are so much better…). For example, Barry Bickmore pointed to Spencer’s elefant-trunk-fitting in one of his books, while Spencer also managed to proclaim that 80% of the CO2 increase was due to ocean warming without thinking about the enormous sink required elsewhere.
Either the guy glows, or he carries a large number of radiative heaters with him…?
The interesting part for me is “this average relationship between temperature trend and population density”. You have to go onto Roy Spencers blog to see the 3 or 4 posts he’s written about that.
If there IS a statistically significant relationship between population density change and rate of warming and that relationship is as high as Roy Spencer shows, then it seems to me unavoidable that he’s right. After-all the warming cannot be causing the population density, it must be the otherway round.
On the otherhand I am quite sure he’s wrong because there are inconsistencies (the UAH trend being one of them). Unfortunately I don’t have the statistical expertize to test it.
One odd thing I thought though was if you go here and scroll down here you come to some scatter graphs with lines of best fit. The R2 values reported are really low. I don’t know much about R2 values but I thought such low values preclude a linear relationship being taken seriously, yet Roy Spencer seems to just base conclusions on them anyway.
http://www.drroyspencer.com/2012/04/ushcn-surface-temperatures-1973-2012-dramatic-warming-adjustments-noisy-trends/
So I think it’s important to test what Roy Spencer did to get the “relationship” between population density and temperature trends because even though on the face of it it makes little sense (the UAH inconsistency pointed out here is but one thing that makes no sense), I am reasonably sure what Roy Spencer claims cannot be true, but I would like to make sure.
The best description I could find of his methodology was this one (which is where I got the above equations):
http://www.drroyspencer.com/2012/04/new-u-s-population-adjusted-temperature-dataset-pdat-1973-2012
where he uses 2000 census population density data (not the density change but just the density) to develop that relationship with temperature trend.
It’s really interesting that when he starts to perform a similar analysis with a different data set (in the post tamino linked) and gets completely different trends, he doesn’t follow through with a new adjustment. If I’m right about how he did his initial adjustment, then he would have gotten a very large trend (+0.46 deg C/decade).
Instead, he develops regional relationships that are all over the map (including several with very large negative relationships), which doesn’t lead him to seriously question his initial hyposthesis. Instead, he just spouts off about the cooling effect of planting crops in arid regions and questions why the USHCN doesn’t correct for this effect. Even though it’s the densely populated cities where this effect is supposedly taking place, not the lowly populated rural areas where you would think most of the planting is occuring. In addition, weaker negative relationships also appear in the hardly arid NW and NE US, but he doesn’t try to explain them.
So, when he tries to apply his physically meaningless curve fitting exercise to a similar data set and doesn’t get the result he wants, he concludes there must be somethng wrong with USHCN rather than his hyposthesis.
How typical.
Nomnomnom: “After-all the warming cannot be causing the population density, it must be the otherway round.”\
Without reference to the data, that statement is not necessarily true. It could be that people have settled in areas that have certain climatic characteristics and those areas happen to be warming faster than others.
Gaz,
“It could be that people have settled in areas that have certain climatic characteristics and those areas happen to be warming faster than others.”
Such an absurd premise that I don’t even know where to begin.
The much more likely explanation for high density urban areas is, of course, quite simply, UHI.
Oops, strike that, as I do know where to begin.
As in follow the WATER!
People have always historically settled where there were/are readily available sources of water, be they wells, streams, ponds, brooks, rivers, lakes or oceans.
Nothing at all to do with climate, everything to do with basic survival. No water, guess what, you die.
There are (a very few) MODERN exceptions to this rule, Las Vagas being a prime example of such.
It would than follow that increased population growth would occur from these core (or root) settlement areas.
Nothing at all to do with “certain climatic characteristics” everything to do with “not dying” or some such.
@Ernst K: it is all because Roy is holding hands with yet another mystery man, John Christy.
This just confirms what many will have suspected all along, Spencer is just disconnected from reality.
Indeed — http://dotearth.blogs.nytimes.com/2012/04/12/on-astronauts-nasa-and-climate-concerns/?comments#permid=49
I asked this question in response to one of his previous posts. His response was as follows:
If that is indeed the 1979-present trend for the U.S. (I haven’t looked), then 4 possibilities come to mind:
1) the difference is real, and surface warming has been compensated by extra convective heat transport from the surface to the troposphere;
2) my analysis is wrong;
3) the UAH LT data over the U.S. are wrong;
4) some combination of the above.
If temperature is being driving by population density, than why are the glaciers in remote areas retreating just as fast as those near populated areas. Take Kerguelen Island, or Reqiang Glacier, Tibet or Rohss Bay, Antarctica or Hornbreen, Svalbard
Mauri, stop asking such rational questions!
Also, there has been a significant loss of population in most rural areas in the Western world which is why using a single census is a major error
That’s a bit like being impressed that more students will sign up for a no-work “Rocks for Jocks” freshman science course than will sign up for
hard-core upper-division geophysics courses.
Well, actually, that’s not a very fair comparison — to the “Rocks for Jocks” class, that is. In a “Rocks for Jocks” class, students will actually *learn* a few things.
And speaking of college, has anybody seen Anthony Watts’ transcript?
Since were are roughly on the topic of the surface temperature record, I thought that I’d toss this out — hopefully, some folks here might find the stuff below useful in swaying “reachable” skeptical friends/family/etc.
Just 3 quick images that I threw together to try to communicate visually to non-technical types just how robust the surface temperature record is (and how bogus Anthony Watts’ claims are).
1) A comparison of the NASA/GISS “meteorological stations” results vs. what I call “Sparse Rural Stations” results (computed from 85 rural sites scattered around the world, an average of 50 of which actually reported data for any month/year). Sites were chosen by dividing the world up into large grid-cells of approximately equal area, selecting the single rural station with the longest temperature record in each grid-cell, and simply averaging all the station/month anomalies together for each year — no area-weighting or anything like that — just plain old averaging! (Really super simple — Tamino would fall asleep from boredom doing something like this.)
Linky here: http://img42.imageshack.us/img42/8018/rural20x20.jpg
Oh yes, and I used *raw*, not “homogenized” data — results were nearly identical for the GHCN V2 and V3 raw data-sets, btw.
2) A Google Earth image with locations of stations used by NASA/GISS.
Linky here: http://img560.imageshack.us/img560/3810/googleearthnasagiss.jpg
3) A Google Earth image with locations of stations used to produce the “sparse rural stations” results in (1).
Linky here: http://img38.imageshack.us/img38/5811/googleearthrural20x20.jpg
There’s absolutely nothing of any real scientific interest here (for folks like Tamino, this is all “been there, done that, wore out the t-shirt” stuff).
But these pix seem to help me drive home to non-scientific types the point that denier claims about UHI, data-adjustments, etc. are completely bogus.
The bottom line here is that I used raw data from small number of rural stations, most of which are in the middle of f*&!ing nowhere (as shown by the Google Earth imagery), ran the raw data through a straightforward averaging process that high-schoolers could understand, and I still ended up getting results that look darned similar to NASA’s.
So what was all that stuff about needing UHI and “data adjustments” to show warming?
You really don’t need to have anything approaching Tamino’s mathematical/technical chops to do this — this is something that really could be broken up into a series of homework assignments appropriate for many computer-savvy students.
Is it any wonder we call them climate deniers?
I am reminded of a comparison of GHCN v2 raw with v2 adjusted data which shows GHCN adjustments make little difference to global warming:
We can see the forest. It’s fine. It’s robust. Deniers ignore that and dive into the trees. They find a couple of bad trees in order to insinuate that the whole forest might be bad.
Can they really be so stupid? On an individual level sure, but collectively, considering the amount of time they’ve poured into the surface record issue over many years, how can they have missed the obvious? No, it has to be deliberate, willful, denial.
nomnom,
Hey, I made that graph way back in the day! They should really update it though, given that it was made back when we thought the v2_adj file was a complete set of stations rather than just those stations that were adjusted (excluding any stations that did not have adjustments applied). Here is a version that correctly treats v2_adj like a change log: http://rankexploits.com/musings/wp-content/uploads/2010/07/Picture-479.png
Thankfully, in GHCN v3 they switched to make both the raw and adjusted files complete.
The opening sentence of my above post got chopped off… so here it is again:
Watts is now chest-thumping about how much more popular WUWT is than are climateprogress, skepticalscience, etc.
With that opening, this sentence should make sense:
That’s a bit like being impressed that more students will sign up for a “Rocks for Jocks” class…
“Watts is now chest-thumping about how much more popular WUWT is than are climateprogress, skepticalscience, etc.”
He always has, I think he has both ego and insecurity issues. He even turned up on my obscure and irrelevant blog when I ripped into a series of his errors to post a childish comment.
For years the most popular newspaper in the UK was the News of the World. That didn’t turn out too well in the end.
Take Kerguelen Island,
OK, let’s take Kerguelen Island. According to Wikipedia, Kerguelen was completely uninhabited until the French established a science research station there a few decades ago. So you’ve gone from a population of zero to 50-100. Plot *that* on a log scale. That’s what’s causing the Kerguelen glaciers to retreat.
You can find similarly large (on a log-scale) population explosions of scientists camped out on glaciers/ice-sheets all over the world. It isn’t global-warming that’s melting all that ice. It’s those scientists, with their heated tents, diesel generators, drilling equipment, etc. who are doing it! It’s the RHI (Research Heat Island) effect.
(If I posted this over at WUWT, how long would it take for someone to pick up this ball and run with it?)
As far as I’ve been able to determine, he went to university (Purdue, IIRC?) for a year or two but did not graduate.
Even with Spencer were right… he forgets that the “urban heat island effect” IS an anthropogenic climate forcing. As in, the temperature change is real, and it is caused by human activity.
Something that has never been really dealt with in the sciencey blogs is that if there is a UHI and if global temperatures are being forced upwards by GHGs, then with increasing urbanization temperatures where people live are going to go through the roof and indeed bodies will be on the pavement (see Chicago/heat wave). It’s not one thing or another, it’s one thing and another.
Tamino,
Thanks for the rather timely post on Woy’s rather bogus analysis.
I’ve tackled Wor’s problem from a number of avenues, Wor’s own statistics from the PDAT analysis and the Columbia University’s 2000 CONUS population (density) dataset(s).
But first a tangent. At the link you point to at WTFUWT? Woy has not posted a single comment since Nick Stokes showed up (to date). Why do you suppose that is so? More to the point where is Woy’s PDAT dataset? I mean why don’t any of the other fake skeptics ask for Woy’s PDAT dataset? You would think that if one were, in fact, a real skeptic, that one would be demanding that Woy release all the data Woy used and whatnot, right?
But no, Woy does not publish the PDAT dataset for others to critically examine in his blog posts. And none of the other fake skeptics have even bothered to ask for Woy’s PDAT dataset AFAIK.
Conclusion? There are no real skeptics at WTFUWT? or at Woy’s blog.
Now on to the PDAT dataset as Woy presents it. Woy starts off by showing a shotgun scatter plot of temperature trends (degrees C/decade) versus population density (raised to the 1/5th power, because, well because how else is Woy going to end up with straight line further on in Woy’s rather bogus analysis) with a magnificant R^2 of just 0.0795.
So with an R^2 of just 0.0795, that value for R^2 would not pass any meaningful test of statistical significance for the null hypothesis of no trend as a function population density (BTW Tamino, please step in and correct me on any of this, as my use of statistical terminology may not be strictly correct).
So what does Woy do? Woy boils down the PDAT dataset to just four data points and fits a linear OLS trend to this redused PDAT dataset. So since a linear OLS has two DOF, that leaves us with just two DOF for the purposes of significance testing of the aforementioned trendline right? Woy now get’s an R^2 of 0.9249 for Woy’s fit through just four data points raised to the 1/5th power. So now, does this trendline show significance to the null hypothesis of no trend line at the p = 0.05 level? And what test of significance should we apply to to such a small dataset of just four datapoints anyways?
Woy than goes on to state: “The standard error of the regression coefficient is +/- 20%, …”
So the trend line starts off at ~0.125 and ends up at ~0.240, so does applying a two sigma test to the this trend line exclude the null hypothesis of no trend line with population density raised to the 1/5th power?
0.125*1.4 = 0.175
0.240*0.6 = 0.144
0.144 < 0.175 meaning that the null hypothesis of no trend line with population density raised to the 1/5th power can not be rejected at the two sigma level.
(Again, I'll gladly stand corrected on any of my misunderstandings on any of this.)
Finally, it does not appear that Woy even bothered with a proper inverse area weighting from Woy's own PDAT dataset.
I'll discuss the Columbia CONUS dataset in a subsequent post, if necessary, but the bottom line there is how much of that data set is rural versus urban populations, and what exactly determines rural versus urban population densities to begin with in the first place (for the purposes of defining significant UHI). And how much of the PDAT dataset is lost if we apply some (admittingly subjective (But how else does one go about defining rural versus urban in an objective manner?)) definition of rural versus urban (remember one sq. mi. = 640 acres).
[Response: I’m afraid some of your statistical statements are off. Perhaps the most important thing to emphasize is that the R^2 value cannot be used to test for statistical significance. You can have high R^2 without significance, or low R^2 with significance.
But from what you’ve said, it sure looks like Spencer’s PDAT is … how shall one say? … something to be skeptical of.]
Hell, they still have not released the code for the TLT series, something folks would really like to audit.
“When “global warming” only shows up after the data are adjusted, one can understand why so many people are suspicious of the adjustments.”
USA =/= The planet.
It seems to me that Spencer should start dropping leaflets from aircraft over all the areas where natural species are altering their migration and other seasonal patterns based on these surface station adjustments. Apparently our nation’s wildlife have not been reading the blogs!
Damm those liberal Pine Bark beetles!
Obviously a bio-weapon created by the Club of Rome in order to socialize forestry in North America…
Hmm, are we going to have to FOIA Roy for his code?
(Kidding, kidding…. surely McIntyre will do that for all us skeptics.)
Someone needs to pull an “Alan Sokal” for WUWT. Especially now that he is trumpeting. Of course, it would be hard for a non-regular to get anything posted, but it would be delicious to post something that delighted deniers (but was obviously and intentionally BS), have them gloat about these wonderful “findings” for a week or so, and then go back a few weeks later and expose the hoax.
Has been done… google “benthic bacteria”
Thanks for that pointer, GP–eventually I found the spoof here:
Click to access Journal_of_Geoclimatic_Studies_2007_23.pdf
For those who, like me, missed it the first time round, it’s well worth a read if you’re in the mood for a bit of foolery. Passages like this:
…led me to consider developing a possible writer’s metric called the BGI (for “Bafflegab Index.”) I’d think that this paper could serve as a useful benchmark for pegging (pun intended) possible maximal values.
From a statistical perspective, the graphs in Figures 1-4 are (and for me at least, literally were) LOLs.
But the “discussion” is the best part…
The USHCN adjusted data show a rise. The raw data show a rise. Spencer’s adjusted data don’t show a rise. Spencer’s adjusted data with the USHCN adjustments show a rise.
If I have that correct, isn’t it possible that the adjustment factors are linked in some way? Perhaps population is linked to time of day adjustment in some way, for example, so that adjusting for population selects the data most in need of time of day correction and most inaccurate without it?
Given that the satellite data and environmental indicators say warming is real, an artefact of such a linkage seems a more plausible explanation to a sceptical mind.
I pointed out to Roy on his blog (http://www.drroyspencer.com/2012/03/mckitrick-michaels-were-right-more-evidence-of-spurious-warming-in-the-ipcc-surface-temperature-dataset/#comment-42550) that his own UAH dataset contradicted this line of reasoning when he first started rehashing M&M. No direct response, but he knows this is an issue as he addresses it in update #2 of his post here http://www.drroyspencer.com/2012/04/new-u-s-population-adjusted-temperature-dataset-pdat-1973-2012/ .
He starts by suggesting there is a step increase in the difference between his surface dataset and his satellite dataset in 1995. If only he knew how to determine if this were statistically significant (my eyecronometer says “no”, but it is his responsibility to test his assertions before he makes them, not mine to do it afterwards). Apparently in coincides with a change in instruments on the satelite data, so I suspect that there will be an update to the satelite data that reduces the warming it shows, making it an even greater outlier.
Spencer really ought to pass his statistical musings before a statistician that actually knows what they are talking about before publishing. All scientists make mistakes, especially if they are at the cutting edge and it is no big deal, the key of course is to find them BEFORE you publish! ;o)
“Nowarming Man”
Horatio Algeranon’s diversion of another classic (Sorry John, RIP)
He’s a real Nowarming man,
Sitting in his Nowarming Land,
Making all his Nowarming plans
for nobody.
Definitely has a point of view,
Knows just where he’s going to,
Isn’t he a bit like John Christy?
Nowarming Man please listen,
You surely know what you’re missing,
Nowarming Man, WUWT is at your command!
He’s as blind as he can be,
Just sees what he wants to see,
Nowarming Man can you see me at all?
Nowarming Man, don’t worry,
Take your time, don’t hurry,
Leave it all till somebody else
lends you a hand!
Definitely has a point of view,
Knows just where he’s going to,
Isn’t he a bit like John Christy?
Nowarming Man please listen,
you surely know what you’re missing
Nowarming Man, WUWT is at your command!
He’s a real Nowarming Man,
Sitting in his Nowarming Land,
Making all his Nowarming plans
for nobody.
Making all his Nowarming plans
for nobody.
Making all his Nowarming plans
for nobody!
Well done. Unfortunately, that means I’m going to be humming the tune all day…
Horatio changed it to “No-warm Man”cuz that sounds better (Horatio has high standards to keep up)
~@:>
Yes, “No-warm” scans much better, and sounds much closer to the original.
I was waiting for “The Magical Mystery Man”..
So many possibilities and so little time.
But if it’s any consolation, Horatio did a parody of another Beatles tune:
“Loopy on the web with deny-mans”
Picture yourself on a blog on the internet
With sciencey themes and American Pi’s
Somebody calls you, you censor quite quickly
A blog with kaleidoscope lies
Surface-station photos of blacktop and grills
Challenging all of the trends
Look for a blogger with the sun in his eyes
And he’s gone
Loopy on the web with deny-mans
Loopy on the web with deny-mans
Loopy on the web with deny-mans, ah
Follow it down to a seat in the Senate
Where oil execs attract George Marshall flies
Everyone smiles as you drift past the dung-heap
That grows so incredibly high
Cosmic ray theories appear on the shore
Waiting to take you away
Go on attack with your head in the clouds
And you’re gone
Loopy on the web with deny-mans
Loopy on the web with deny-mans
Loopy on the web with deny-mans, ah
Picture yourself snapping photos of stations
With FOX News reporters with fossil fuel ties
Suddenly, someone is there with a term style
The blog with kaleidoscope lies
Loopy on the web with deny-mans
Loopy on the web with deny-mans
Loopy on the web with deny-mans, ah
Loopy on the web with deny-mans
Loopy on the web with deny-mans
Loopy on the web with deny-mans, ah
Loopy on the web with deny-mans
Loopy on the web with deny-mans
Loopy on the web with deny-mans
I like these lines:
“Everyone smiles as you drift past the dung-heap
That grows so incredibly high”
OK, call me a vulgarian.
Yes, well Horatio always liked the German phrase “Haufen Mist”, which evokes the image of a ripe, steaming pile of manure
It’s actually part of a children’s song:
Ist das nicht ein ByBESTdissed
Ja, das nicht ein Haufen Mist
sorry, should be
Ist das nicht ein ByBESTdissed?
Ja, das ist ein Haufen Mist
Thank you, that was indeed some consolati0n…
Claims of cooling are not going away, despite the evidence. The fact that the nonsense below is still on-line, and quoted as fact by some Canadian deniers, may indicate that Spencer’s convoluted logic (and worse statistics) will form a lasting citation of solid science for the denial movement.
quote:
Frontier Centre: We are all familiar with the modern theory that the world’s climate is getting warmer. Is it?
Tim Ball: Yes, it warmed from 1680 up to 1940, but since 1940 it’s been cooling down. The evidence for warming is because of distorted records. The satellite data, for example, shows cooling.
FC: Could you summarize the evidence that suggests the world is cooling slightly, not warming up?
TB: Yes, since 1940 and from 1940 until 1980, even the surface record shows cooling. The argument is that there has been warming since then but, in fact, almost all of that is due to what is called the “urban heat island” effect – that is, that the weather stations are around the edge of cities and the cities expanded out and distorted the record. When you look at rural stations – if you look at the Antarctic, for example – the South Pole shows cooling since 1957 and the satellite data which has been up since 1978 shows a slight cooling trend as well.
end quote:
http://www.fcpp.org/publication.php/864
Hmmph. Ball should cue in the other professional dissemblers:
“Everybody agrees that the temperature has warmed. The people who disagree about temperatures are the barking mad end of the spectrum.”
David Whitehouse
21 October 2011
[Source]
“All sceptics believe in “global warming” (depending on what time scale you use); what they doubt to various degrees is the “man made” element.”
James Delingpole
21 October 2011
[Source]
“You’ve all seen articles say that global warming stopped in 1998. Well, with all due respect, that’s being a little bit unfair to the data…it was a huge El Niño year, and the sun was very active in 1998…make an argument that you can get killed on, and you will kill us [skeptics] all..if you lose credibility on this issue, you lose the issue.”
Patrick Michaels
6 September 2009
[Source]
and
“what’s happened – and this is why this [global cooling] argument is so very very dangerous – is that solar activity and the El Niño or La Niña we’re in now have conspired to add up to produce very very little temperature change in the last couple of years…and so what’s going to happen, one of these years, that’s going to turn around. And if you make that argument now, you’re going to have a very difficult time defending the future.”
Patrick Michaels
6 September 2009
[Source]
I never realized that the professional septics actually gave a monkey’s about credibility or making consistent arguments.
Hell, 2008 was meant to be the start of a big cooling trend.
They were relying of HadCRU3 to show ‘no warming since 199X’ even whilst claiming that ClimateGate completely discredited HadCRU3..
“It warmed up from 1680 up to 1940, but since 1940 it’s been cooling down…”
Now, there’s a piece of insightful, trustworthy analysis… maybe his ‘code’ should be requested, too?
Of course, Ball also claimed that Andrew Weaver was ‘backing away’ from the IPCC–Dr. Weaver is only lead author for Chapter 12 of the forthcoming AR5–among other (allegedly*) false and libelous assertions. Oh, and made some rather, er, “inflated” claims as to his own credentials.
Clearly Ball is an unreliable source, to say the least.
*Dr. Weaver’s suit is still before the Canadian courts. Personally, I expect Ball’s nose to be thoroughly bloodied, legally speaking, of course.
I noticed another silliness about Roy’s analysis. He regresses the temp trend against PD^0.2, and then says that the trend thus “predicted” can be subtracted from the measured trend.,
But that corrects to zero PD. And that’s not the situation in America. In fact, the PD of ConUS is about 40/sq km. And if you read that value off his regression, it corresponds to about 0.12 C/decade.
On that basis his PDA corrected value would be 0.133C/decade, not 0.013.
Actually, from Woy’s PDAT dataset (N = 280, a number divisible by four), I’d SWAG about 160 people/km^2 (M = 160^0.2 = 2.76, for 40 people/km^2 that gives M = 40^0.2 = 2.09, which from Woy’s graph would only include seven stations total) from the upper graph from this figure (I counted 266 points myself and then estimated the median value by again counting 133 points along the x-axis, with the expectation that the median isn’t that much different from the mean);
So using 160 people/km^2 (instead of 40 people/km^2);
0.0587 * 2.76 + 0,0128 = 0.175 degrees C/decade
I’ve assumed that the 0.013 you’ve mentioned above is the 0.0128 value Woy gets from his rather curious regression (mind you no PDAT dataset raw data are available and no error bars presented anywhere to even begin to gage statistical significance).
If one chooses the lower graph from the aforementioned figure, with it’s grand total of four data points (and assuming that each of those data points represents 70 data points from the upper figure), then the mean is ~2.8^5 = 172 people/km^2 (So we can reasonably conclude that the mean of that PDAT dataset is somewhere between say 2.7^5 and 2.8^5).
So using 172 people/km^2 (instead of 40 people/km^2);
0.0586 * 2.8 + 0,0128 = 0.177 degrees C/decade
Lessons learned?
Well Woy thinks that he can remove both the mean value and the trendline slope from the raw data without presenting any statistical tests of significance for the trendline slope. But this would be like setting the CONUS mean warming to 0 degrees C/decade.
Now why do you suppose that Woy would want to do that?
Horatio,
LOL! That was one of your better ones. Not that there are any bad ones… Er., I mean… I’ll shut up now.
EFS Junior,
“Such an absurd premise that I don’t even know where to begin.”
No, actually, it’s not absurd.
And I think you’re missing my point.
If indeed the areas where people settled had a common characteristic of ready availablity of water as you suggest (or transport routes, or agricultural productivity, or nice views whatever other features that made them attractive for settlement) it is entirely plausible that such characteristics may have been associated with certain climatic characteristics (including a greater or lesser propensity to warm due to AGW).
I think my comment about Nomnomnom’s remark (Nom said: “After-all the warming cannot be causing the population density, it must be the otherway round.”) should have been completely uncontroversial.
Just because two things are significantly correlated does not mean that either A must cause B or B must cause A. They could just as well both be caused by C.
Just a matter of logic. As I said, it’s not a comment on actual conditions in those places.
One of Roy’s first blog posts explaining a method of looking at temps vs population density involved using some method of estimating areas and applying a correction to temps based on elevation. His amazing results were that temps changed most strongly for population changes at the very low end (i.e. from 1 to a few/kmsq) and then the slope decreased sharply from that lower rate for higher populations. This makes no physical sense.
What struck me at the time was that there was likely a problem in the elevation correction for temps, and that this was strongly correlated to increasing population density. Low population densities in any area, would generally be the higher elevations, with increasing densities at lower elevations, and the relationship might well look like the very nonlinear result he got. He used a constant term for correction vs elevation.
Someone smarter than me in stats needs to look at that issue.
Yes, his 1/5 power law leads to odd results. A large part of his low trend (at least 0.05C/decade) happens as the temp trend is corrected right down from PD 1/km2 to zero, Odd things happen with his rule near zero PD. I calculated that, if the US really were depopulated, then the entry of just one inhabitant would raise the ConUS trend by 0.0023 C/decade.
Dr. Spencer claimed ignorance regarding the stark trend difference with the UAH LT trend on his blog today (5/5/2012). No explanation. Claimed he didn’t even calculate or consider the UAH LT trend. Both disappointing and seemingly not very credible.
Ah, the “incompetence” defense.