HadCrut3: October Warmer than September.
The HadCrut3 anomaly for October is out. The most recent release indicates October’s anomaly is 0.440 C; this represents a rise since September’s current Hadcrut3 anomaly of 0.371C.
Figure 1: Hadley October Warmer than September.
Other changes in recent temperatures include:
- September got colder during October, dropping from 0.376 to 0.371. In contrast at GISS, September got 0.1 C warmer during October.
- August got colder during October, dropping from 0.385C to 0.384C. This change is insignificangt compared to the 0.11C drop in August’s temperature in the GISSTemp record.
- July warmed up during October, rising from 0.412C to 0.414C. Recall that GISSTemp July dropped 0.1C during October.
So, if we ignore changes in the third significant figure, Hadley’s temperatures were stable after 1 month. My general impression is that’s the way Hadley’s temperatures behave. (But the point of initiating these sorts of comparisons last month was to see if my impressions were correct. People had been wondering.)
Brownie betting on GISSTemps’s October Anomaly are now closed! So, I’ll tabulate and we can see who wins.
Comments
lucia
(Comment#6676)
November 14th, 2008 at 10:46 am
Nick– This does not clear GISS of failure to include the most rudimentary of automated checks, but it explains why there is no particular problem with earlier months. The current months anoalie always changes a little the following month. This is because not all data are in and small mistakes are probably made. That’s not what happened this month at GISS.
NOAA contributed to this months error. GISS’s failure to catch errors only matters when NOAA makes a mistake.
Keith (Comment#6680)
November 14th, 2008 at 11:47 am
Well, with this news, Lucia, I’m pretty sure that GISSTemp will stay around the 0.52-0.54 tamp range. They’ve got their support group from Hadley, so they can stay in the upper ranges. Now, by December, there may be some settling downward by both sets as the final temps come in from around the world. But I doubt it drops more than 0.05 degrees.
Nick (Comment#6685)
November 14th, 2008 at 12:30 pm
NOAA contributed to this months error. GISS’s failure to catch errors only matters when NOAA makes a mistake. This does not clear GISS of failure to include the most rudimentary of automated checks, but it explains why there is no particular problem with earlier months. That just pushes the problem back one stage to NOAA. NOAA didn’t catch it. So their filters aren’t set up to catch the problem. That means they should check their raw data and adjust if necessary. GISS can then redo all their data and adjust back. The logic that applies to GISS not checking, applies to NOAA not checking.
Arthur Smith (Comment#6688)
November 14th, 2008 at 12:42 pm
Lucia - in Hadley numbers, September only dropped 0.005 degrees, from 0.376 to 0.371, not 0.385 to 0.371. 0.385 was the old August number. At least from your copy of the table. By the way, is anybody still keeping track of the betting from early this year on what the year-end GISS land-ocean average anomaly would be for 2008? As I recall, your number (from “lumpy”) was 0.70 +- 0.11 C: http://rankexploits.com/musing.....rediction/ and various guesses collected here: http://www.climateaudit.org/?p=2698 My guess was 0.41 C for 2008, though I had the advantage of being another month later in the year (see my Feb 15 comment). With ten months of data, the GISS average so far this year is now 0.405 C, while the average going back 12 months (including Nov and Dec 2007) is 0.411 C. Still 2 months of real data to go, but I’m thinking my chances are pretty good!
steven mosher (Comment#6690)
November 14th, 2008 at 1:27 pm
Thanks for finding that arthur, NOT! my guess of .23C is looking
pretty lame. Although I did offer even money on an over/under bet at .38C
lucia
(Comment#6694)
November 14th, 2008 at 2:23 pm
Arthur– Maybe I’ll have to re-troactively promise brownies. I wasn’t doing that then, but getting the whole year right is pretty good. I may also need to re-do Lumpy. (I had a few thoughts. Probably won’t improve her any, but she’ll change.)
I’ll have to fish those out! I was way high, wasn’t I? I think David Stockwell collected bets too.
lucia
(Comment#6695)
November 14th, 2008 at 2:31 pm
Arthur– Thanks. I fixed the number in the post. I wonder to what extent that even ordinary circumstances, greater stability of Hadley numbers is purely a function of waiting just a few days? One thing we learned from the October Anomaly Kerfuffle is that the Nov 10 number were based on about 1/2 the stations. By Nov. 13 a few more had come it. It may be that Hadley waits until some minimum fraction of stations report and GISS tries to push their reports out on the 10th. (Those dang Americans and their milestones!)
steven mosher (Comment#6703)
November 14th, 2008 at 5:52 pm
lucia if you look at the CA thread ( arthur linked) you’ll find
a bunch of guesses made in Jan 08 9 and some in feb. in one comment I collated a bunch. Also I think there were some made on WUWT, but I cant find the thread.
Mike C (Comment#6716)
November 14th, 2008 at 10:11 pm
Lucia, I’m suprised that you are comparing GISS and HADCRUT without discussing high lattitude temperatures since that is the major difference between the two datasets. Keep that up and the AGW boogie man will start calling you an unsufruct gorilla while having a legitimate gripe.
lucia
(Comment#6718)
November 14th, 2008 at 10:14 pm
MikeC– Just reporting the recently release monthly values! Do you think the large revisions in already published numbers at GISS is due to their interpolation to estimate the temperatures over the arctic where no one has a thermometer?
lucia
(Comment#6719)
November 14th, 2008 at 10:16 pm
What’s an “unsufruct gorilla?”
Mike Bryant (Comment#6720)
November 14th, 2008 at 10:32 pm
I think it’s a gorilla suit that you do not own, however you can use it if you don’t mess it up.
Mike C (Comment#6721)
November 14th, 2008 at 11:13 pm
Lucia, if you know not the unsufruct gorilla when it comes to Climate Auditors catching GISS with their pants down, then you are waaaaaaaaaaaay blogging the wrong subject. As for your other question, “Do you think the large revisions in already published numbers at GISS is due to their interpolation to estimate the temperatures over the arctic where no one has a thermometer?”
History lesson here:
http://www.climateaudit.org/?p=1880
and here:
http://www.columbia.edu/~jeh1/.....aldeal.pdf
This question is either ridiculously rhetorical or indicative of the hostess being completely in the green zone(using the inexperienced paradigm). You cannot make a comparrison unless you remove the high lattitudes from Hansen. It is why his temps tend to be warmer as of recent, and cooler the farther back ya go. So to get a solid comparrison between the two, you have to remove Hansen’s Arctic and Antarctic extrapolations. After removing the aforementioned guesses, then your 3 “Other changes in recent temperatures include:” statement would make scientific sense.
MartinGAtkins (Comment#6726)
November 15th, 2008 at 2:31 am
I tried re-editing my last message three times. Here is the corrected post. Lucia “The HadCrut3 anomaly for October is out. The most recent release indicates October’s anomaly is 0.440 C; this represents a rise since September’s current Hadcrut3 anomaly of 0.371C” http://www.remss.com/data/msu/monthly_time_series/ TLT MSU Global TLT MSU USA TLT MSU North Hem. TLT= Lower Troposphere. I don’t have the numbers for Siberia. ;-}
2008 9 0.194
2008 10 0.181
Colder.
2008 9 0.25
2008 10 0.135
Colder
2008 9 0.228
2008 10 0.283
Warmer
lucia
(Comment#6731)
November 15th, 2008 at 7:48 am
MikeC– It got caught in Spam. In this post, I only compare the two in the sense that both claim to provide products that measure the same thing (GMST) and that one is more variable from month to month. That’s why I asked if you thought the variations are due to the pole business. I know that GISS’s extrapolation takes the exact same measurements and weights the ones near the pole to extrapolate over that area, claiming that’s better. Hadley doesn’t. So, that makes a difference. I wasn’t familiar with the usufrunct & the gorilla title. Yep. GISS made a mistake in processing, it existed a long time, it practically took an act of Congress to get them to correct and admit the error, and then Hansen put out a newsletter explaining how unimportant mistakes in GISS are. I suspect this is what Boris referred to as “Hansen Derangement Syndrome”?
Mike C (Comment#6733)
November 15th, 2008 at 9:40 am
Actually, it did not take an act of congress, Hansen corrected the problem quickly and initially noted Steve Mc for catching the problem. The response by the bloggosphere and talk radio (all the way to Rush Limbaugh) to the error was enormous.
“Hansen Derangement Syndrome” would be that Skeptics hate Hansen so much that they are going to rake him over the coals about anything. It’s like Bush Derangement Syndrome; some people hate him so much that he could come on TV warning of an asteroid or something and Bush haters would find 10 minutes to stand in front of their TV to scream obsenities at him instead of taking cover (to use an example from the Red State Update comedy show).
lucia
(Comment#6735)
November 15th, 2008 at 10:37 am
I stand corrected. Hansen fixed quickly when the error was found On HDS: You are probably correct. At this point, there are some who would lambaste Hansen for tieing his shoes the wrong way. On the other hand, Hansen is doing a pretty good job providing things for people to complain about. ( The analogy to Bush is apt!)
Sekerob (Comment#6745)
November 15th, 2008 at 12:07 pm
You guys are not moving fast enough. NCDC published theirs and it’s up, plus several prior months were adjusted… for late data i suppose, but feel free to issue a pre-emptive FOI to demand explanations for the “unresolved discrepancies” Oddly, if i fit the MSU data and the GISS/HadCRU/NCDC data, but from baseline shifts there’s barely a difference of meaning. Who knows do they check off each others results. cheerio oh yes, the last 12 months 2008 1 0.1959
2008 2 0.3695
2008 3 0.7179
2008 4 0.4296
2008 5 0.4402
2008 6 0.4829
2008 7 0.5066
2008 8 0.4658
2008 9 0.4438
2008 10 0.6307
John M (Comment#6750)
November 15th, 2008 at 1:49 pm
Sekerob (Comment#6745) I doubt an FOI will be required. Delta anomaly, Oct 07 to Oct 08 GISS: +0.05 Delta anomaly, Sep 08 to Oct 08 GISS: +0.08 Disclaimer, I may have missed some adjustments somewhere along the way.
HadCRUT: +0.073
NOAA: +0.134
RSS: -0.044
UAH: -0.064
HadCRUT: +0.069
NOAA: +0.187
RSS: -0.013
UAH: +0.006
MartinGAtkins (Comment#6773)
November 16th, 2008 at 10:19 am
October 2008 SST Update For those that get high on information overload, Bob Tisdale has updated his graphs.
OPENING NOTES: All but the graph of weekly NINO3.4 SST anomalies are based on ERSST.v2 data. They illustrate raw monthly data from January 1978 to October 2008. The graph of weekly NINO3.4 SST anomalies is based on OI.v2 SST data and illustrates raw weekly data that’s centered on Wednesdays.
Bob B (Comment#6798)
November 17th, 2008 at 8:29 am
The surface station data are all crap and should be discarded.
Fred Nieuwenhuis
(Comment#6799)
November 17th, 2008 at 8:55 am
I notice that NOAA has delayed the release of their Global analysis (here: http://www.ncdc.noaa.gov/oa/cl.....troduction) They originally were going to release it last week. Hmmmm..
Fred Nieuwenhuis
(Comment#6800)
November 17th, 2008 at 8:58 am
This story has been picked in a major daily in Canada:
http://www.nationalpost.com/op.....f13fd759ec
lucia
(Comment#6803)
November 17th, 2008 at 9:14 am
It’s funny to read the stories in the news. I’d say the Canadian article did a better job than the earlier UK one. Still… is this true: This October wasn’t the warmest October ever, it was only the 70th-warmest in the past 114 years — that is, in the bottom half of all Octobers, not at the top of the list. So why the massive discrepancy between the published GISS numbers and the correct ones? I guess I’ll have to go make a histogram. I have the numbers. . .
lucia
(Comment#6805)
November 17th, 2008 at 9:38 am
I get that based on GISS, the anomaly of 55/100 C makes October the 5th warmest October since 1880. It’s tied with four other moths for 49-53rd warmest anomalie since 1977 (or if you prefer 330 coolest.) Maybe the Canadian article is using another temperature service? (If so, they should say so.) October wasn’t cool.
Fred Nieuwenhuis
(Comment#6809)
November 17th, 2008 at 10:48 am
I have emailed the article’s author. Let’s see what comes out.
Steve McIntyre
(Comment#6810)
November 17th, 2008 at 11:14 am
I’ve emailed him as well, notifying him of a problem.
Fred Nieuwenhuis
(Comment#6811)
November 17th, 2008 at 12:22 pm
Received this response: “We making a correction for tomorrow. It was the 70th warmest in the US, not worldwide. Worldwide, depending on who numbers one uses, it was somewhere between 5th and 8th.”
lucia
(Comment#6812)
November 17th, 2008 at 12:25 pm
Gosh… why do the wait until tomorrow? At least for the online journal, you’d think they would have the ability to revise today. Tomorrow the article will have be quoted all over the place. Sigh. . .
lucia
(Comment#6813)
November 17th, 2008 at 1:00 pm
I emailed too. I guess I should have thought to do that when I first noticed. They might have it changed by now! (Or not.)
MartinGAtkins (Comment#6822)
November 18th, 2008 at 2:01 am
For those that get high on information overload, Bob Tisdale has updated his graphs. Bob B(Comment#6798) November 17th, 2008 at 8:29 am I agree with you. It appears that the SST graphs used by Bob are contaminated with surface station data. I use RSS/MSU data and none show a dramatic rise in northern hemisphere ocean temps for October. Going through the data set compilations it appears that surface station data is used to estimate SST. Don’t ask me why.
http://bobtisdale.blogspot.com/
The surface station data are all crap and should be discarded.
Fred Nieuwenhuis
(Comment#6825)
November 18th, 2008 at 9:13 am
Here’s the correction, a lousy one. http://www.nationalpost.com/story.html?id=967585 PS. NOAA still hasn’t updated their Global Analysis page. Hmmmm.
lucia
(Comment#6826)
November 18th, 2008 at 9:22 am
Fred– I guess unlike bloggers, online papers don’t run their corrections on the page with the article itself? Or credit how the boo-boo came to their attention? How is anyone supposed to find that? Sigh. . . Lorne Gunter (the author) emailed me. It appear he doesn’t have control over posting corrections. I guess that’s newspapers for you!
Fred Nieuwenhuis
(Comment#6828)
November 18th, 2008 at 9:44 am
By the way, I noticed that there is still issues October GISTemp with northern and western Canada. As of this morning, the without SST data and 250 km smooth, virtually all of Canada is blank. However when the SST analysis is added (and 250 km smoothing), all kinds of colours pop up. The most interesting is the land locked purple patch. See here:
http://x9c.xanga.com/a32807f26.....514822.jpg
http://xe8.xanga.com/09782bea3.....514821.jpg
Fred Nieuwenhuis
(Comment#6829)
November 18th, 2008 at 9:47 am
Ooops that didn’t work right. Try these:
http://x9c.xanga.com/a32807f26.....514822.jpg and
http://xe8.xanga.com/09782bea3.....514821.jpg
lucia
(Comment#6832)
November 18th, 2008 at 11:04 am
Fred– Heck, for all we know, the services will suddenly discover… something… and the we will suddenly discover the world has been warming since 2001!
Yep. Values are slow this month. Whatever NOAA went wrong at NOAA is finally being realized at GISS, and things are slow! For the time being, we can all only wonder if the current issue is related to the long standing issues Steve M has been blogging about.
BarryW (Comment#6834)
November 18th, 2008 at 2:10 pm
Fred Nieuwenhuis (Comment#6829) Something seems really wrong. Look at Alaska, purple next to brown ( +4) in both graphs.
Mike Bryant (Comment#6841)
November 18th, 2008 at 4:35 pm
It looks like the wheels are coming off the wagon.
MartinGAtkins (Comment#6850)
November 19th, 2008 at 11:45 am
BarryW (Comment#6834) November 18th, 2008 at 2:10 pm Something seems really wrong. Look at Alaska, purple next to brown ( +4) in both graphs. It’s worse if you consider you have purple contrasted next to deep blue in the south. Barrow station is probably one of the longest serving in that area and it’s data is in for October. I ran a preliminary plot on the October numbers and they do show it as being warm over the entire data set. I’ll tidy things up a little and see if i can find an unusual trend jump. Meanwhile the data is here :-
http://data.giss.nasa.gov/work.....tation.txt
Fred Nieuwenhuis
(Comment#6851)
November 19th, 2008 at 12:15 pm
I am willing to buy the Barrow (Alaskan) data/maps. Barrow is on the north shore of Alaska which was ice free for the majority of October. Therefore any northerly winds would be modified by the open water, giving higher than normal temps. Fairbanks in the middle of Alaska, would be cooled by those very same northerly winds and therefore would have lower than normal temps. According to what data I have been able to find, the October ‘08 mean temp for Barrow was -5.1 C and the mean for Fairbanks was -9.0 C. Depending what the “normal” is for each location (I couldn’t find this data) would give a significant spread of anomaly values. However, the big purple patch in the middle of land when SST analysis is added to the map is the one that gets me. Does the SST analysis take into account LAKE temperatures???
BarryW (Comment#6856)
November 19th, 2008 at 7:38 pm
Point of reference on Barrow. Over at CA Steve is looking at NOAA data and found some Oct 2008 outliers. http://www.climateaudit.org/?p=4395#comment-312590 He’s looking specifically at FInland, but one of the sites mentioned was Barrow. Don’t know if it has the same weirdness as Steve is seeing with the Finnish sites or not.
Fred Nieuwenhuis
(Comment#6870)
November 21st, 2008 at 8:51 am
They (whoever they is) has “fixed” the issue with the Northern Canada anomalies mentioned above by removing data. Now most of Canada is blank. Which is a problem all on its own as the Canadian data IS available. Here is the new map:
http://x55.xanga.com/80ef3207c.....850382.gif
BarryW (Comment#6871)
November 21st, 2008 at 9:50 am
So then we have “darkest Africa” and now “darkest Canada”. Or grayest, eh?
Fred Nieuwenhuis
(Comment#6873)
November 21st, 2008 at 10:32 am
Second largest country but good area coverage with stations, one of the best and most readily available climate datasets in the world and it’s GREY.
Bunch of hosers, eh.
lucia
(Comment#6874)
November 21st, 2008 at 10:35 am
Fred
If we decide to see this as the glass is half full, the vanishing stations means someone is doing work and they are fixing something.
Lucia
Fred Nieuwenhuis
(Comment#6876)
November 21st, 2008 at 10:55 am
I dunno…maybe…the people at NOAA are doing something anyway. http://www.climateaudit.org/?p=4414
?>
Nick (Comment#6674) November 14th, 2008 at 10:19 am
1. GISS didn’t pick up on the error.
2. That means that GISS data filters did not check for this type of error.
3. If they didn’t check last month, they didn’t check the months before.
4. They only changed the recent data, not the past data.
That’s particularly strange and needs an explaination. Why is it that their filters only work on the last month, but not prior months? You would expect at least some of the historical months to have similar errors, and so their new filter for bad values should have picked them up. That would mean that prior months temperatures would be adjusted (up or down!).
They have only adjusted the recent, and that needs some explaination