By Steve Goddard
Dr. John Christy recently wrote an excellent piece “Is Jim Hansen’s Global Temperature Skillful?” which highlighted how poorly Dr. Hansen’s past predictions are doing.
This post raises questions about GISS claims of record 2010 temperatures. The most recent GISS graph below shows nearly constant warming from 1965 to the present, with 2010 almost 0.1°C warmer than the actual warmest year of 1998.
HadCrut disagrees. They show temperatures flat over the past decade. and 2010 about 0.1°C cooler than the warmest year 1998.
Looking more closely, the normalised plot below shows trends from Jan 1998 to the present for GISS, HadCrut, UAH and RSS
GISS shows much more warming than anybody else during that period. Hansen claims :
The difference of +0.08°C compared with 2005, the prior warmest year, is large enough that 2010 is likely, but not certain, to be the warmest year in the GISS record.
The discrepancy with the other data sources is larger than Hansen’s claimed 0.08 record. Is it a record temperature, or is it good old fashioned bad data?
Either way, it is still far below Hansen’s projected temperatures for 2010. This is not pretty science.
Hansen made temperature forecasts which have proven too high. Now his “measured” temperature data is pushing higher than everyone else. Would you accept the other team’s coach doing double duty as the referee? In what other profession would people accept this sort of conflict of interest?
The idea that James Hansen has doctored the data is a totally non-falsifiable conspiracy theory. I notice you don’t accuse anyone in the the Hadley Center or the CRU of adjusting the HadCRUT data downwards to make it more flat, even though that is just as likely (read: just as ridiculous).
The GISS graph presented is NOT a calendar year time series. The graph is a 12-month running mean running from mid-year to mid-year. GISS does not pretend otherwise.
The point of that graph is to show what the *last* 12 months look like in comparison to previous twelvemonth, mid-year to mid-year periods. Obviously, this will result in changes for the whole record.
While it’s obviously a good convention, there is no mathematical reason why analysis of yearly global temps should be run on a Jan – Dec basis. Think of the tax year, for example.
Steve, could you please label the graph more clearly, so that it is not confused with the calendar year GISS record, or with the other calendar year records?
Re HadCRUt: your assertion that 2010 is currently looking cooler than 1998 is wrong. The graph you link to only goes up to 2009. Here is the link to the HadCRUt graph with the current year included (based on an average of the months so far).
http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/
Would you please amend that, too?
barry
HadCrut January to June 1998 – 0.58
HadCrut January to June 2010 – 0.52
This includes HadCrut’s recent upwards (cough cough) adjustments.
July/August, 1998 were very high, so it is likely that 2010 will drop further relative to 1998. Hope this helps.
stevengoddard, for what it’s worth I share your suspicions. I can’t reinforce your suspicions because I depend on information from you in forming my opinion. I don’t have sufficient expertise. But, bottom line, grave suspicion is distinct from clear evidence and definitive proof. I think you’ve got clear evidence of poor or questionable practice. The problem for me is the leap from evidence of questionable practice, which you’ve presented, to definitive proof of motive, which I don’t think can be established. That doesn’t mean we can’t suspect it, but that’s different.
Leone
DMI showed record low (1958-2002) temperatures north of 80N.
GISS showed above average (1958-2002) temperatures north of 80N.
You don’t see a problem there?
Leone says: “Two things to consider:
“1. When looking recent temp trends, one shall reduce Pinatubo away from data before conclusions. Here is the graph showing HadCRUT timeseries with Pinatubo reduced…”
Have you also adjusted for other volcanic eruptions such as El Chiochon?
You continued, “Used correction values are obtained from Hansen’s research…”
You should check to see if the Hadley Centre has published volcanic aerosol adjustments, because the year-to-year variability of GISTEMP and HADCRUT are different. It would make a difference.
Steve Goddard wrote: “Hansen admits that GISS diverges from HadCrut since 1998, because he claims better Arctic coverage.”
We all understand that the GISS creates the illusion of a complete Arctic dataset, but GISS does have better land surface data coverage in the Arctic; that is, they use more surface stations than Hadley Centre.
Also, if you would, please add the two TLT datasets to all of you follow-up trend comparisons with the different start dates. The trend graphs you’ve provided in the comments don’t have the same datasets as the trend comparison you included in the post, and someone might think you’re trying to hide the increases in the TLT trends when you changed start years.
” Steven mosher says:
August 16, 2010 at 10:40 pm
If, 80N 60E sees .5C of warming since 1980 and if 80N, -120E also sees .5C of warming do you figure that 90,0 saw?
1. the same warming (Giss)
2. Less warming ( evidence?)
3. More warming ( latest re analysis of Sat. data)?”
Asking questions like these only makes sense if we do NOT have any reliable data. But we do. A lot of measuring is going on in the arctic (even if there is no stationary station on the pole due to obvious reasons), and the DMI publishes trustworthy temperatures. So I say with confidence that GISS cheats.
(“Trustworthy”: DMI is in the warmers camp. So if their data does not confirm warming, it must be because they are honest.)
” Leone says:
August 17, 2010 at 5:20 am
While methods are different, it is still statistically strange that GISS seems always show warmer.”
That would be OK, but they show more warmING, which means that the difference gets bigger with time.
DMI are meteorologist, they believe temperature should be measured with thermometers and similar things. GISS are climatologists, they believe temperature is to be estimated from a great big distance.
Steven Goddard: Regarding my earlier comment, here, I’ll save you some time.
The following is a comparison graph from January 1998 to May 2010, with linear trends, including GISS, Hadley Centre, and NCDC surface temperature and RSS and UAH TLT data. As you’ll note, the GISS and NCDC trends are highest, with NCDC a distant second. But also note how the two TLT datasets have larger responses to the 1997/98 El Niño than the three surface-based datasets. The two TLT datasets also have a greater lag, meaning they stayed at the elevated levels longer into 1998.
http://i36.tinypic.com/23m9c7k.jpg
And now if you start the trend comparison in 1999, the linear trends UAH and RSS TLT anomalies are now much closer to the GISS trend, because they overreact to the La Niña.
http://i33.tinypic.com/28md8iq.jpg
This is one of the hazards of short-term linear trend comparisons. No matter what you select for a start date, someone is always going to comment about it.
It doesn’t. You say:
Then you display a graph from HadCRUt that only goes to 2009, claiming it is showing us 2010 temperatures.
Here is the current HadCRUt graph, including 2010 up to June (soon to include July).
http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/
Bob,
The amplification of satellite temps is exactly the reason why you have to do a peak-to-peak (1998-2010) comparison to get meaningful results vs surface data. Starting in 1999 is not reasonable.
No surprise that GISS and NCDC are similar. Their data sets are not independent.
Weird. I have not discussed “suspicion,” “evil” or anything of the sort.
Pointing out the appearance of a conflict of interest is not an indictment.
I can not be responsible for what goes on in other people’s heads.
The only thing that matters is data. It’s hard to argue with data Steven Goddard is presenting. GISS is different. And his point about ENSO amplification in satellite when comparing it to GISS seems to be getting overlooked.
It looks like trend is a distraction to temperature. Arctic trends are distracting from the actual temperature shown by GISS. I don’t care whet their anomaly shows. I do care about a map that shows red there when in reality it should be blue.
I posted this comment http://wattsupwiththat.com/2010/08/16/is-hansens-recent-temperature-data-consistent/#comment-459182 only in response to people bringing up motives and also because the idea floating around that if something criminal is going on in global warming we should just call it standard practice, forget about it, and go back to business as usual.
This post was not about intent but became about that.
If satellite peaks are amplifications of surface peaks, then the comparison is distorted by the amplification. I.E, larger peaks will have a bigger difference in temps than smaller ones. In this case, 1997/98 was a super el Nino, and 2009/10 only a moderate one. This will skew the results when comparing surface and satellite records.
A much better method is to use longer time series, where the signal to noise ratio is greater. 13 years is not long enough – it fails statistical significance precisely because the data is too noisy to compute meaningful trends over that time period. Amplified peaks and troughs only compound the problem for comparisons.
Just once, I would like to see someone take a Global Temperature Dataset 18** – 2010, and plot it from zero on the Y-axis.
It is getting very old looking at rubber-banded Y-Axes.
barry
If I measure mid-point to mid-point people complain. If I measure peak-to-peak people complain.
For some reason, people seem to be fine with Hansen’s trough to peak, which is the most meaningless of all possible measurements. Even worse, trough to peak error gets worse in satellite data because of amplification.
The only measurement which makes any sense in this case is peak to peak i.e. 1998 to 2010.
barry
Given that we haven’t completed 2010, it is going to be pretty difficult to provide 2010 annual data isn’t it.
Warmists want to switch over to July-July data before La Nina kicks in. i.e. catch the full El Nino cycle for maximum warming. LOL
I can’t imagine why our friends are so eager to finalize 2010 data – before temperatures start tanking from la Nina.
Some of the most transparent cherry-picking ever. Lame.
barry says:
August 17, 2010 at 8:05 am
‘The amplification of satellite temps is exactly the reason why you have to do a peak-to-peak (1998-2010) comparison to get meaningful results vs surface data.’
‘A much better method is to use longer time series, where the signal to noise ratio is greater. 13 years is not long enough – it fails statistical significance precisely because the data is too noisy to compute meaningful trends over that time period. Amplified peaks and troughs only compound the problem for comparisons.’
Right, 12 years (peak 1998 to peak 2010) only give the trend for these 12 years, and there is neither warming or cooling. But 30 years are not very significant, either (UAH satellite temperatures start in Dezember 1978). Now, just what are we to do if we want to compare 1979-1998 with 1998-2010 to see if there is a change?
Steven Mosher said, “Hansen’s method is a bit quirky here and there, but unlike many people who blather on about it I actually read all the damn code back in 2007. Go to CA and look at the threads. I’ve also read all of the python refactor and I run the code on my machine. ”
Steven, who you recall how UHI adjustment is implemented in GISTEMP? Do you recall discussing UHI implementation in the CA thread(s) you referred to earlier?
Dave Yup.
I have issues with the Adjustment that have evolved over time. Go figure. You get to educate yourself over time. Basically this:
1. The criteria for urban ( nightlights) is not the best proxy. Ron Broberg, (mostly) Zeke and I have been working on different criteria. We all have different approaches to the problem. It will be interesting to see what MATERIAL difference that makes
2. My preferred approach is just to drop Urban. No adjustment. Live with the spatial uncertainty.
3. Hansen’s adjustment is just a meat grinder that may get some right and some wrong. There are cases where it looks flat wrong. They bear some investigation, but in the end, the adjustment doesnt amount to much. CERTAINLY not all the warming in the record. Looking at unadjusted data shows you that. looking at just rural shows you that. looking at satillite data shows you that.
4. For grins I will probably do my own adjustment just to prove the point.
stevemosher.wordpress.com
“The only thing that matters is data. It’s hard to argue with data Steven Goddard is presenting. GISS is different.”
yes, and if it was EXACTLY the same you would say whats the point of having CRU and GISS. It’s different. Is it materially different? depends upon your purpose. A BMW is materially different than a Ford. but for the purpose of going to the store both will do. As spock would say a difference that makes no difference, makes no difference. if you have a different purpose then of course the GISS difference may make a difference. Purpose matters. and so, it becomes a discussion of intent, whether intended or not