Update: Robot Change Detection alerted me to a second GISSTemp update. The reported August 2009 anomaly rose to 0.52C, up from 0.51C on Tuesday! A whole bunch of data must have arrived at GISS because there were quite a few changes propagating back to October 1882. Below, I’ve inserted a screen shot of updated temperatures for this century:
The changes are all small, but it’s always a bit amazing to see how far they propagate backwards. It’s also a bit unusual to get a second update in an individual month.
You can compare the cached versions at robot change detection.
Why does anyone pay attention to GISS again??????
The changes are small, but month-by-month over many years probably add up. I work in a company where the performance of employees in a group is compared to the other employees in the group. I have joked before that to get a good review, I don’t have to rise up, I only need to push them down (a better metaphor for my actual thoughts on this is that a rising tide lifts all boats, I like the group I’m in to outperform others if possible, but I digress). It appears that GISS works like this, modern temperatures don’t need to rise to show warming if historical temperatures slowly get adjusted down.
James H (Comment#19718)
The way we put it is that you don’t have to outrun the bear, you just have to outrun the guy running next to you.
JamesH–
I’m not sure the changes consistently make the past colder. At least sometimes past temperature have risen!
That said, the shifting past temperature tends to give ‘some’ the heebee-jeebees.
kuhnkat: You asked, “Why does anyone pay attention to GISS again??????”
Actually, I enjoy downloading the GISTEMP data for specific areas of the globe using KNMI’s Climate Explorer, and then comparing them to the output of the GISS Model E, also available on Climate Explorer, to see just how poorly the Model E represents the GISS real world data. It’s enlightening. Lots of taxpayer money spent with not much to say for it.
And to think the iPCC plans to use GCMs to make regional projections.
> A whole bunch of data must have arrived at GISS
> because there were quite a few changes propagating
> back to October 1882.
My understanding is that it has something to do with the GISS estimation algorithm for missing data. A sane algorithm would…
1) look at surrounding data for that month
2) compare the temperatures over several years when the value was present
3) using 1) and 2) above, generate an estimated value, and stick to it
Jim Hansen’s algorithm use *PERIOD OF RECORD* for the comparisons/adjustments. Of course, as more data is added to the record, numbers change. Isn’t there a rule that you shouldn’t use different populations for comparisons/adjustment? If we accept that the climate has changed over the past century and a quarter, how can we justify using 2009 data to estimate 1882 data?
Interesting example of “refine the ‘Garbage In’ so as to get the desired ‘Garbage Out’ “. Par for the course for GISS, I’m afraid.
Upon this “haute cuisine” GIGO, governments everywhere continue to concoct billion dollar “climate change abatement” programs which they think will please the voters and get them re-elected or maintain them in power, as the political case may be. As fate would have it, the good folks on the streets are increasingly not buying in, as evidenced by what has been happening in Australia, Canada and the steadily growing backlash in the EU.
Walter Dnes,
Does this mean that creeping urbanization will increase the UHI effect. Imagine if there is a site (Site A) that is urban and surrounded by urban sites and so, all else being equal, is warming faster than it was 10 years ago when it was rural and surrounded by rural sites i.e. there is the UHI effect on top of the general warming trend. Now in the past (say 1890) this Site A was obviously rural and surrounded by rural sites. If in say Aug 1890 Site A sent in no data this data must be estimated. 10 years of increasing urbanization has meant a increasing warming rate for Site A so when this is backcast it results in a lower estimate of Aug 1890 than would have been made in 1999 before urbanization. DO I have this correct?
Are there any actual stations left in the GISSTemp setup or do machines running models and correction algorithms just communicate and tabulate once a month about what the data should be?
Interestingly, years ago NASA was challenged by “skeptics” (actually, nutters) about the alleged doctoring of pictures taken by space probes. Some people were accusing the Agency of removing evidence of past Martian civilizations, for example.
Without much fuss, and without any childish posturing and whining, NASA changed its policy so that anybody and everybody can get the “raw” images as soon as they arrive from the probe.
And of course there is no change of past data with any sort of sleight-of-hand hush-hushed data modification.
The fact that _that_ still happens in climate-related NASA circles should make everybody think.
This is simply underscoring why the AGW promotion industry should be given zero credibility at all.
If this was financial data in a public company being treated the way GISS treats temperatures, people would be getting fired, sued or indicted.
Is there any credible way to pose that current claims are valid if the past data sets are rewritten to make those claims?
Bob Tisdale,
” I enjoy downloading the GISTEMP data for specific areas of the globe using KNMI’s Climate Explorer, and then comparing them to the output of the GISS Model E, also available on Climate Explorer,…”
So, this is your monthly dose of climate humor??
kuhnkat: Humor? You think I’m joking? When I was a young man, I had many other interests. Now, not so many. Picking apart the Model E projections are one of the things I do when I get tired of finding El Nino-induced step changes in SST anomalies…
http://bobtisdale.blogspot.com/2009/01/can-el-nino-events-explain-all-of.html
http://bobtisdale.blogspot.com/2009/01/can-el-nino-events-explain-all-of_11.html
…and in TLT anomalies:
http://bobtisdale.blogspot.com/2009/06/rss-msu-tlt-time-latitude-plots.html
…and in OHC anomalies:
http://bobtisdale.blogspot.com/2009/09/enso-dominates-nodc-ocean-heat-content.html
Maybe when you get to be my age…
Bob,
“Maybe when you get to be my age…”
True, I’m only 57.
Actually I didn’t think you were joking. I thought you found the comparisons humorous!!
“whole bunch of data must have arrived at GISS because there were quite a few changes propagating back to October 1882.”
How do you know it’s due to more data? Maybe they threw some out!
Hey, stranger things have happened.
Could it just be changes to a few stations inside the base period? That would affect the whole record.
Aslak–
Yes. It could be that. I don’t know why they should change, but with GISS that would do it.
Hadcrut is more stable.
The changes were already up by September the 11th, which is when I downloaded this month’s dataset for the first time. I would say that the unusual thing was the first dataset being published so early. They hardly ever publish anything before 10th, and 14th or 15th of the month are typical. Even this update was published quite early, for GISS standards. I would think that the previous version had still little data for the month, and that’s the reason of the update.
This said, here is a review of the changes in past records, from the august version.
1) Many months warmed, not so many cooled. In total 78 months (5.02% of the total monthly records) showed some change, although 73 of them only changed 0.01C. Most of the changes happened before 1950 (warming records) and after 1998 (mixed warming and cooling records). One month in particular, January 2006, suffered a surprisingly big change by cooling 0.04C.
2) The global trend cooled a little bit, as well as the trends for the periods 1910-1940, 1941-1970, 1971-2000 and 2001-2008.
3) 2002 average temperature anomaly rised to +0.56C, and now ties with 1998 and 2007 as 2nd to 4th warmest years. Warmest is still 2005 with +0.62C.
4) The temperature anomaly from december 1999 to november 2000 rose +0.01C, although, month by month inside that period, we see 2 months cooling 0.01C and only one warming 0.01C. This is because GISS uses more decimal positions for the temperature anomalies than they publish. There is probably some generalised warming that is not enough to show up in the individual months but do change their average anomaly for the whole period.
Regards.