What sudden recent warming? What Hockey Stick? I don’t see any.
By Lucy Skywalker Green World Trust
with thanks to the late John Daly and his timeless, brilliant website page “What the Stations Say” (click on Arctic map above). Click on each thumbnail graph to access Daly’s full size graph with time and temperature scales and other details. The thicker dark horizontal line across some of these thumbnails indicated 0ºC (a few of the graphs are ALL under that line). The Arctic is shown in the condition of summer sea ice (see thumbnail below) and the pale circle is the Arctic Circle. All data comes from NASA GISS or CRU originally.
Paul Vaughan notes at WUWT that he “spent a fair amount of time updating these graphs (& others of Daly’s for other regions)” using http://data.giss.nasa.gov/gistemp/station_data/ and adds a cautionary note: The time-frame and aspect-ratio of the timeplots can be manipulated to create the illusion of a steep trend in recent years.
The highly variable temperatures and amounts of sea ice in both polar regions is well-known to locals, but cherrypicked extremes have become a media weapon to scare ignorant folk with. Greenlanders today are aware of recent warming; but history, archaeology, and the Norse sagas show that Greenland was warmer than today in the Middle Ages, when crops and trees were grown there. For recent sea ice changes (since 1979) see Cryosphere Today and note that while Northern Hemisphere sea ice (at the top of the CT page) has gone down recently (but is currently going up again), Southern Hemisphere sea ice (at the bottom of the CT page) is going up, so that the overall total is pretty constant although fluctuating between summer and winter.
Finally, Jeff Id’s superb animation of recent Arctic sea ice>>
|
Nick Stokes (16:32:45) :
Frank K: Hi Nick – could you describe in detail how GISS does their zonal averaging and homogenization for the Arctic?
It’s described in detail
http://www.ncdc.noaa.gov/oa/climate/research/ushcn/#dataaccess
here, and in the cited papers. Homogenization modifies sites that are out of line with near neighbours, and so has rare application in the Arctic. It also has little effect on the zonal average, which is the measure I recommend.
Sorry Nick, but the question was about the GIStemp “homogenization” not the USHCN testing of homogenization (i.e. SHAP) which is what your link is about. USHCN only tangentially enters GIStemp in STEP0 (where it throws out most of it and blends the rest with the GHCN copy via a strange method).
The GIStemp process is “strange and wonderous” to say the least. The first thing you must answer is “which homogenization” since it does it in several steps. The one most folks think about is in PApars.f which you can inspect here:
http://chiefio.wordpress.com/2009/03/08/gistemp-step2_papars/
This is part of STEP2 that does the zonalizing.
Luckily this code is modestly well commented
C**** Combining of rural stations
C**** ===========================
C**** Stations within Rngbr km of the urban center U contribute
C**** to the mean at U with weight 1.- d/Rngbr (d = distance
C**** between rural and urban station in km). To remove the station
C**** bias, station data are shifted before combining them with the
C**** current mean. The shift is such that the means over the time
C**** period they have in common remains unchanged. If that common
C**** period is less than 20(NCRIT) years, the station is disregarded.
C**** To decrease that chance, stations are combined successively in
C**** order of the length of their time record.
C**** The homogeneity adjustment parameters
C**** =====================================
C**** To minimize the impact of the natural local variability, only
C**** that part of the combined rural record is actually used that is
C**** supported by at least 3 stations, i.e. heads and tails of the
C**** record that are based on only 1 or 2 stations are dropped. The
C**** difference between that truncated combination and the non-rural
C**** record is found and the best linear fit and best fit by a broken
C**** line (with a variable “knee”) to that difference series are found.
C**** The parameters defining those 2 approximations are tabulated.
C****
C**** Note: No attempt is made to find the longterm trends for urban
C**** and rural combination separately; using the difference only
C**** minimizes the impact of short term regional events that
C**** affect both rural and urban stations, hence cancel out.
What this says is that a semi-random collection of of “nearby” stations up to 1000 km away are all ‘averaged together’ with an attempt to to remove ‘station bias’ via a semi-random overlap interval used as a stable mean that the other curve is adjusted to match. Then any ends with less than three stations in them are pruned off.
THEN this average is used to “adjust” the target station.
Does any of this make sense? I don’t think so. It “cools” the history of Pisa by 1.4C and if you exclude Alpine stations from the adjusting set (i.e. stations over 900m) the bogus cooling of the past reduces to 1.0C.
Oh, and while all the 2 through N stations have their “Bias Removed” via this mean comparison, station one gets used “as is where is” (and, IMHO, creates a bias by setting the mean to itself. Yes, I think this is a bug, but it needs confirmation.) See:
http://chiefio.wordpress.com/2009/08/30/gistemp-a-slice-of-pisa/
Almost the same code is used in STEP3 to “homogenize” the grids and boxes.
Scott Mandia (17:44:49) :
So you can explain the Roman & Medieval warm periods and associated following cooling rather than trying to just disappear them?
If so then you are the one for the Nobel science, (not peace), prize.
Just because you don’t understand what the cause is, is not a proof of non-existence.
The only reason for the various hokey (correct spelling) sticks has been to do this disappearing act so the current warming, (probably now cooling), appears extraordinary.
Incidentally. What is the unit of forcing?
DaveE.
I thought we were warming since the LIA. Why was the Artic cooling from then until mid 2000’s according to the Kaufman paper? Isn’t this just as disconcerting as the purported warming for the last 50 years?
The ‘hockey stick’ shape refers to millenial reconstructions, yet the top post here discusses temps for a century and less, weirdly conflating two distinct periods (no one claims the instrumental record has a hockey stick shape).
No trend analysis on the stations on the stations mentioned above, a large nunmber of stations left out, data needlessly truncated (there is more data for most of the stations above – why not include it?), and virtually all the stations mentioned show a warming trend anyway.
I call cherry-picking and distortion. When someone gathers the data for all Arctic stations (north of L60?), and plots all data, not just summertime, then we’ll have something with the beginnings of robustness. This eyeballing a few hand-picked time series with gargantuan, trend-hiding Y axes, doesn’t even come close.
Should have said – “all Arctic stations with sufficient time/data”.
Adam Grey (20:49:35) : “The ‘hockey stick’ shape refers to millenial reconstructions, yet the top post here discusses temps for a century and less, weirdly conflating two distinct periods (no one claims the instrumental record has a hockey stick shape).”
Huh???? Wha???
The hockey stick is clearly and expressly shown in a century or less….
You are trying to compare apples….to green beans….and it ain’t flyin!
Weirdly conflating what??? Huh???
DUH.
Chris
Norfolk, VA, USA
Notice how he uses all the sound-byte “power phrases”:
“Conflating”
“Eyeballing”
“Truncating”
“Trend-Analysis”
“Cherry-picking”
http://www.thefreedictionary.com/sophistry
Chris
Norfolk, VA, USA
Adam Grey (20:49:35) “[…] large nunmber of stations left out, data needlessly truncated (there is more data for most of the stations above – why not include it?), […] I call cherry-picking and distortion.”
–
Nick Stokes (01:55:48) “[…] cherry-pick […]”
—
You have misunderstood.
Clarification:
1) John Daly is dead (so he cannot update his work).
2) Not all of the Arctic stations have long records.
What Lucy has done:
a) reorganized Daly’s work into an intuitive visual format.
b) indicated a desire to collaborate to update the graphs.
Seems like a large majority of the proxies in Kaufmann aren’t even correlated to temperature. A good percentage seem to be used to cancel the LIA/MWP (inverted from what you’d expect) or add slope . Shouldn’t each proxy be shown to correlate to temperature before being averaged to imply temperature? I mean if they don’t show a LIA or MWP (yes, you can claim I expect to see this, at least in the northern hemisphere) or especially if they show the opposite trend, I would expect them to require validation as having a temperature signal and define polarity of the proxy.
Just try grouping each subset together and see what story they tell…even common proxies don’t tell the same story. What a mess. Seems like there are lags as well in some proxies. I don’t see how you could ever get a temperature signal out of that combination of proxies…
Looks to me that you could pick your specific combination of proxies to achieve the desired trend.
Hmmm…seems like a good career choice.
Ed
Thanks EM Smith. Nick does not get the mess that GISSTemp is. you have had more patience with the code than I have, giving up a few months after hansen released it. It really annoys me when people trot out GISSTEMP series like it was a record of observations. It’s not. There is even MORE fun if you jump down the rabbit hole of the GISSTEMP data SOURCE! GHCN and then jump down THAT rabbit hole to the B91s.
I’ll be glad when you get a version up and running. There are tons of sensitivity experiments.
RR Kampen (06:09:00) :
do you think certain AGW proponents like Al Gore have no financial interests?
Adam Grey (20:49:35) :
The ‘hockey stick’ shape refers to millenial reconstructions, yet the top post here discusses temps for a century and less, weirdly conflating two distinct periods (no one claims the instrumental record has a hockey stick shape).
No trend analysis on the stations on the stations mentioned above, a large nunmber of stations left out, data needlessly truncated (there is more data for most of the stations above – why not include it?), and virtually all the stations mentioned show a warming trend anyway.
I call cherry-picking and distortion. When someone gathers the data for all Arctic stations (north of L60?), and plots all data, not just summertime, then we’ll have something with the beginnings of robustness. This eyeballing a few hand-picked time series with gargantuan, trend-hiding Y axes, doesn’t even come close.
And what trend was officially found for temps in the 20th century? 0.7 degrees Celcius. REF: http://epa.gov/climatechange/science/recenttc_triad.html
And what were the error bounds for the instruments used to establish that trend in the 20th Century? The USHCN Network has 91% of it’s instruments with an error bar greater than 1 degree celcius.
REF: http://www.surfacestations.org/
How do you know that the claimed 0.7 degree Celcius up trend actually occured when so many instruments can’t measure to that level of accuracy?
I would also challenge you to demonstrate a pre 20th Century temperature proxy that has an error band less than 1 degree celcius.
From where I’m standing, I don’t see how you can justify a conclusion that you know that there is a rising temperature trend given the instruments and proxy data that you have to work with.
savethesharks,
It’s been a long time since I played hockey, but if I remember correctly the stick has a long handle with a short foot, something like this:
_____________/
The term is attributed to millenial constructions which approximate that shape, as the regulars here well know.
The instrumental record doesn’t look like that, whether global or Arctic. More like rising sine wave with a flat middle. The term ‘hockey stick’ in this context is bewildering (unless it’s an orthognal jab at another topic).
I’m not sure what you mean by ‘power phrases’. The terms you listed are used by people here, including Anthony Watts. Do you contend they are also engaging in sophistry or does a different standard apply?
John Vaughan – thanks for the clarification. I had read that you updated the data, and so wondered why there was nothing more recent than 2003 and many series finished earlier. Even though this is so far a collative work, I still hold that a better reflection of the graphs would be all Y axes that show variabiliy (and trend) more clearly, up-to-date and all stations (as caveated in my previous post) – just for eyeballing purposes. A statistical analysis with as much servicable data as possible would be much better. Has such been done anywhere that you know of (not just summer temps)? Apart from the science institutes that collate the data, I mean.
As the sites posting this information in this way have a clear agenda – nothing wrong with that per se – I hope you can understand that I am deeply skeptical of selective presentations. (Yes, I apply my skepticism equally)
Tamino’s reply to this post is worth checking out.
Martin Mason (11:35:03) :
Imagine that. People are susceptible to corruption and fraud.
And some people are scientists, which means…
Paul Vaughan (14:05:22) :
“Espen, I suggest you check your claim.”
Sorry about that, I was too sloppy in making that claim because I wanted to make my point about the rapidly diminishing number of reporting arctic stations.
In another post, you asked for official data from Norway. You can get forecasts and recent data from yr.no (also in english) and history and trends from eklima.met.no (free, but registration required), but I think the data at rimfrost.no should be the same.
Peter Taylor (02:57:37) :The Greenland temperatures are perhaps the most important, and these are the key ones showing the latest peak – but there is a difference between east and west Greenland. If you look at the Arctic sea ice animations, this observation makes sense – ocean currents you can actually see at work.
I have come to the conclusion from this Arctic data that if greenhouse gases are having an effect (since 1940), then it can’t be more than about 20% – and the same conclusion can be made from satellite data of the increased short-wave flux to the global oceans from 1980-2000 caused by thinning cloud – the net radiation gain is about four times that computed for infra-red ‘radiative forcing’. I’d like to know how you quantify the 20% – another day!
As you may be aware, Syun-Ichi Akasofu, formerly head of the International Arctic Research Centre in Fairbanks, thinks the overall trend from 1800 hasn’t changed much as a ‘recovery from the Little Ice Age’ I know and like his work. However, I think he trends the recovery too eagerly as an overall straight line rise. I think it’s far more likely to be cycles-within-cycles, with the longterm recovery from LIA slowing down slowly.
Sean Ogilvie (04:44:52) :… As has been demonstrated here there is no accurate long term data. GISS claims to go back to 1880 but that is a joke. I think NASA was better when Daly first used it. And please note, these are individual, mainly-rural stations that have not been moved and “upgraded” degraded, with many records taken in times when your son’s life might depend on the record’s accuracy.
Let’s look at the quality control of GISS. The worst I’ve seen recently is an anomaly of +19.4967 C (+35.0941 F) near the city of Akzigit Kazakhstan for January 2002. Is this the material that was here a while back, when all the brouhaha about the Russian record hiatuses blew up? It would be brilliant to see a post that covers the GISS worldwide shortfalls so you can see the full picture: the worst and the average / pattern of distortions made by the degraded records.
Bill Illis (05:42:38) : There is one proxy which drives the big blade on the hockey stick in this Kaufman Arctic study – Series 22 – which rises to a huge +6.97C in the most recent decades. That is Briffa’s Yamal Pennisula tree ring reconstruction (This is the same one that Steve McIntyre has been trying to get the data and methods for but has not been successful.)
Thanks Bill, I think it would make a good post here if you could write up a progress report from CA plus the essence of the Yamal story. I really appreciated your link to the pdf showing that the original Yamal material has no hockey stick at all – if I understand you right. Now I want to know what Briffa did to it.
Scott,
I couldn’t AGREE with you more!
I doubt it. If you look at my post carefully, I am just saying more or less the opposite of what you seem to believe.
In a nutshell: “AGW is not an issue. Mankind is not facing destruction caused by CO2. Rather, resources that would be used more effectively elsewhere are being wasted. Polititians and environmental groups are not helping us, nor can they by their very nature”.
I suspect you don’t agree with the above. Specially since I believe yourself are one of these resources that are being wasted.
And that is a pity because, although I obviously think you are mistaken, I admire your ability to discuss inteligently without losing your temper. You deserve to be enrolled in a better cause, and not in the one that is bringing a lot of money and power to some smart guys.
I notice that giss has severe gaps in icelandic temperatures for a data update an analysis see this icelandic source:
http://en.vedur.is/climatology/clim/nr/1213
in particular:
summer temperatures
http://en.vedur.is/media/loftslag/myndasafn/medium/loftslag-hiti4.png
winter temperatures
http://en.vedur.is/media/loftslag/myndasafn/medium/loftslag-hiti3.png
ref:
Hanna, E., T.Jónsson, J.E.Box (2004): An analysis of Icelandic climate since the nineteenth century. International J. of Climatology 24, p. 1193-2004.
savethesharks (21:35:51) “Trend-Analysis”
I also found such calls amusing.
I live in a jurisdiction that has a carbon tax. I’ve seen reports from a “climate institute” (financed by the government here) that would suggest:
a) night-time lows are going to overtake day-time highs.
b) Winter temperatures are going to overtake summer temperatures.
A bright Stat 101 student will easily point out that the model assumptions are not met – (and so trend analysis is garbage in this context).
I have just come across a very interesting post on Climate Sanity that works on an animation presented here:
http://climatesanity.wordpress.com/2009/09/09/dmi-arctic-temperature-data-does-show-increasing-temperature-trend/
The animation seems to show, as I have understood it, that one of the biggest lies in that AGW proponents have told is that the recent warming period (which I believe has really happened), is not such a threat.
The global mean temperature did probably increase in the last years of the 20th Century, but not mainly because days in the summer in the lower latitudes were hotter, but in part because nights in the winter had become less severe in the higher latitudes.
How that is bad?
Oliver Ramsay (07:18:15) : …Several times last winter I tried to alert the people at wunderground to the curious fact that the forecasts for these two locations each day were in the -30C range (just like the rest of the Territory) but the recorded temperatures were consistently in the range of +5C. I received one e-mail in acknowledgement with an assurance that it would be checked out. Nothing changed and I heard no more. Seeing Nick’s citing of warming in Fort Simpson, which is a proximal station, I am wondering again about the smearing of data that Surface Stations has described. It’s my understanding that the source of the Canadian data is Environment Canada, but their own site didn’t seem to list these particular stations. Does anybody know about this?
Looks like there’s room for another whole piece here. I’m more and more sure that the whole AGW can be put to rest on the multiple faulting of the most basic records. Can you dig some more?
Espen (07:19:58) : Juraj V: The rimfrost.no site that Nick recommended also shows much more moderate trends if you choose the 100 year trend option. And if they had offered a 70-year trend option, we’d see almost zero trend for most station…
Hei Espen, takk for alt du har skrivet her. Har du sitt, Jo Novas bok er oversettet på norsk?
Can you write up all this into another piece? I could add it to our website. Vær så god, email meg.
Ferdinand Engelbeen (09:13:35) : Nice work lucy! A few years ago I searched for the (raw) data of all circumpolar stations over 66.6N with sufficiently long records. About 70% were warmer to equal warm in the 1935-1950 period than in the current period. 30% are warmer in the current period, mainly in Eastern Siberia and Alaska. The same for Greenland, I plotted all stations around the inland ice up to 2005 here…
Thanks Ferdinand! Your link is not working however and I’d love to see your work which seems like a backup to this piece…
Adam Grey (00:18:07) “Tamino’s reply to this post is worth checking out.”
I see Tamino is quoting me severely out-of-context. An honest mistake?…
–
You asked about analyses. I’ve no doubt you’ll find no shortage of analyses in the literature. Also, I encourage you to run your own analyses. One of my interests is the timing of major turning points in climate records.
–
It is interesting to see how eager some are to distort my comments. Thanks for the alert Adam.
Espen (00:18:40) “I wanted to make my point about the rapidly diminishing number of reporting arctic stations.”
I agree that this is an unwelcome development.
–
Thank you for the info on accessing Norwegian climate series. My hope is to study them in depth.
Scott Mandia (07:11:43) : …Of course, proxy data is not as high quality as actual measurements but in studies of climate we do not have the luxury of direct measurements for the past. At some point, you have to accept the proxy data with a certain degree of error otherwise we might as well just not study climate change. My point still stands that camparing plots that are 100 years or fewer and that end in 2002 or earlier, CANNOT be used to debunk a study that shows a 2,000 year trend has changed in the past 100 years and even more so in the previous ten years.
Scott, I remember how it was when I believed in AGW. The way I drew on arguments, “evidence”, and “proof”, and the way I viewed certain prominent skeptical contributors, was, I now realize, highly coloured by the fact that I believed the basic “warmist” scientific arguments, and therefore tended to accept more of their material without question. I can recognize the same energy behind everything you write, it’s a very subtle effect a lot of the time, but it is nevertheless powerful in causing even fair evidence to be misunderstood, and I don’t think it will change unless I can show you that every single one of those pieces on which you draw from Gristmill and RealClimate (as well as comparable pieces from New Scientist, BBC, etc) is cardinally flawed. It took me a long time to deconstruct those pieces to my own satisfaction – and it would take even longer to deconstruct them to the satisfaction of others who lapse into belief more readily than I – I’ve always been a terrier for digging out and checking evidence.
Still, I’ll have a go.
proxy data is not as high quality as actual measurements but in studies of climate we do not have the luxury of direct measurements for the past As I said, I recognize the style of language here. Of course, proxy data is not as high quality as actual measurements. This is one reason why I focussed on direct measurements and left the proxies aside. Measurements whose accuracy, in 1860, might well make the difference between your son’s life and death. Now while we do not appear to have “direct measurements” from the Arctic earlier than 1820, we do have other indicators beside the proxies now so widely touted. And those other indicators are distinguished by the deafening silence with which they are not mentioned in the current science – as if they are even less respectable than proxies. Now this is a point that needs examination, not just sliding over. History and archaeology show that Greenland was inhabited, and fields were cultivated, in the Middle Ages. Some of the sites still lie in the current permafrost – yet there is no way they could have been permafrost-bitten originally. I believe there is even evidence of trees, though I cannot confirm this firsthand. There is a huge, utterly huge reservoir of such studies. The true longterm patterns of glaciers. Massive evidence that the Medieval Warm Period was warmer than today. I have not got the patience to cite all the references here now. Some are in my basic piece (click on my name) or in other pages on our website. Much is yielded by careful Googling.
At some point, you have to accept the proxy data with a certain degree of error otherwise we might as well just not study climate change. What Climate Audit have found is that the level of error accepted by Mann et al (“the Team”) is not acceptable – and that certain methodology is highly slipshod (it would be failed in any engineering-quality report) – and that there are plenty of studies with far less error that show the Medieval Warm Period etc – and that the unacceptable errors ALL have to do with the difference between alarmist levels of warming, and no warming at all, or only insignificant / natural warming. Again, you need to read something basic here like Caspar and the Jesus Paper
My point still stands that camparing plots that are 100 years or fewer and that end in 2002 or earlier, CANNOT be used to debunk a study that shows a 2,000 year trend has changed in the past 100 years and even more so in the previous ten years. I had a close look at the latest “hockey stick” and it clearly starts rising sharply in, or just before, 1900. Therefore the WHOLE BLADE is within the range of actual temperature measurements. This is already enough evidence to show that the purported blade does not appear to exist at all in the good long records. It may well exist in the records from stations that suffer UHI or have been moved and upgraded put in range of warm effects like planes, maritime moderation, etc – Anthony is the expert on how many different modern measurement factors can each cause a warm bias. Lastly, the tiny difference caused by the fact that Daly died in 2004 does not change the massive import of the records from 1820 on up to 2003 or so. I’ve checked some recent NASA records, and some do indeed appear to show a little uptick at the end of the record. But I know that NASA now use “homogenization” techniques that were not used earlier, which often means that good rural records are distorted by bad station records. And since I see gross spiking in earlier records, and a tame tiny uptick in recent ones, my BS detector has been activated. No proof, and no time to prove it right now. So I leave it aside since my reasoning does NOT stand or fall on five years’ evidence.
But I have noted it.
One thing which occurs to me after staring at all these graphs and temperature lists from the arctic, is that the period that GISS uses as base period for its anomaly graphs almost coincides with a cold cycle in the Arctic. If they had used an earlier period, or a longer period, the numbers would have looked pretty different. For instance, the 1951-1980 mean annual temperature for Angmassalik, Greenland is -1.27, but the1921 – 1980 mean annual temperature is -0.88. I picked another station with long-running data, Haparanda in Finland, and computed 0.98 for the 1951-1980 period and for 1921-1980 1.29. Or how about 1.6 for the 1921-1950 period?