Here is a review of the major global temperature metrics in tabular and graph form. There is a bit of disagreement this month. GISS still comes out the warmest, as it did last month, and the month before, and there is a bit of divergence between the RSS and UAH satelitte derived datasets.
I’m a little late to this game as I’ve been busy catching up on personal business since my trip to NCDC Asheville and 20 station survey across North Carolina, but I thought it was worth a review.
RSS (Remote Sensing Systems of Santa Rosa, CA) RSS Microwave Sounder Unit (MSU) global temperature anomaly data by For April 2008 has moved a tiny bit higher, with a value of .080°C for a change (∆T) of 0.001°C globally from March.
RSS
2008 1 -0.070
2008 2 -0.002
2008 3 0.079
2008 4 0.080
click image for a ZOOMED 1998 -2008 DATA PORTION of the 1979-2008 image
RAW RSS data is available here
Note that there does not appear to be any sustained upwards trend post 1998.
Here is the entire RSS MSU dataset plotted:
click for a larger image
Reference: RSS data here (RSS Data Version 3.1)
University of Alabama, Huntsville (UAH) published their data set for April 2008, and it is in slight disagreement with the RSS data, dropping with greater magnitude than RSS. The ∆T for April 2008 was -.079°C for an anomaly value of 0.015°C
UAH
2008 1 -0.046
2008 2 0.020
2008 3 0.094
2008 4 0.015
Click for a larger image
Reference: UAH lower troposphere data
Both RSS and UAH datasets are just slightly about the zero anomaly line, which when compared to the last peak in global temperature in January 2007, is about 0.5°C cooler. That drop in the last year seems to be holding for now.
HadCRUT (Hadley Climate Research Unit Temperature, UK) surface, land-ocean
Last month HadCRUT took a positive jump to 0.430°C but went down again this month in agreement with drops noted in RSS and UAH global datasets. This month’s anomaly is 0.250°C for a change (∆T) of -0.18°C globally from March 2008.
2008/01 0.056
2008/02 0.187
2008/03 0.430
2008/04 0.250
Click for a larger image
Reference: HadCRUT3 anomaly data which can be found here
description of the HadCRUT3 data file columns is here
GISS (Goddard Institute for Space Studies, New York)
(surface, land-ocean, polar estimates)
Year Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2007 .86 .63 .60 .64 .56 .53 .53 .57 .51 .55 .49 .40
2008 .12 .26 .67 .41
2008 .13 .26 .60 .41
I don’t bother plotting GISS global temperature anomaly data anymore for two reasons:
1) The format that the data is presented in requires me to do extra work to get it into an easily plotable form since it spans rows and columns, instead of being one long linear column of data for easy import.
2) I simply don’t trust the representivity of the GISS dataset. Since GISS uses polar “interpolated” data which does not exist in any of the other three datasets, I see GISS as an outlier and not truly representative of global temperature. For example, GISS interpolates from the closest high latitude surface stations and adds that data for the very high latitude arctic where no weather stations exist. None of the other 3 data sets use interpolation to create data for the arctic where no measurements exist. Hence, I have a greater degree of trust for RSS, UAH, and HadCRUT global temperature anomaly data.
And there’s more, see a post called “The Accidental Tourist” about GISS data problems, from John Goetz (originally posted on Climate Audit) which I’ve reposted as the next item down.
Reference: GISS dataset temperature index data here
In summary, it appears that all four global datasets are maintaining the drop in temperature noted globally in the last year since the last peak in January 2007. With the exception of GISS, the other three datasets show an anomaly just slightly above the zero baseline for April 2008.
Click for an interactive graph
UPDATE2: Thanks to Barry Hearn at http://www.junkscience.com and his processed CSV file I was able to easily plot the GISS Global Temperature anomaly without going through formatting conversions, the result is shown below:
Click for a larger image
I noted that unlike the other three metrics, GISS shows peaks in 2002 and 2007 that exceed the peak of 1998 in magnitude. How can GISS arrive at these values with combined land+ocean when HadCRUT does not? Curious.
Dear G-d, we’re all going to die…eventually.
The surface station outdoes the sats. Again. How very unexpected.
Is there ANY easy way to get data in a plotable form: I mean, the HADCrut data is in 12 columns and the only way I’ve managed to plot that is to paste it into Word and painstakingly Enter, click, Enter, click etc until it’s in one column. Same with other data with multiple columns where you’ve got to delete loads. Am I missing something: I mean what makes GISS so much harder to plot than the rest? Two columns is low compared with the others.
How are the 12-month running means to the end of April 2008 doing? Here are some comparisons. Note about my numbers…
1987.917 = November 1987
1988.000 = December 1987
1988.083 = January 1988
etc, etc.
It looks like
– Hadley is cooler than 12 month mean ending November 1997.
– GISS is cooler than 12 month mean ending January 1998.
– UAH is cooler than 12 month mean ending August 1988.
– RSS is cooler than 12 month mean ending September 1988.
For a plot, see http://www.woodfortrees.org/plot/rss/mean:12/from:1979/plot/uah/mean:12/from:1979/plot/hadcrut3gl/mean:12/from:1979/plot/gistemp/mean:12/from:1979
Here’s hoping the code tags work…
The most recent 12 month running means
Date Hadley GISS UAH RSS
2007.417 0.375 0.560 0.199 0.232
2007.500 0.376 0.530 0.203 0.280
2007.583 0.403 0.530 0.255 0.363
2007.667 0.370 0.570 0.286 0.367
2007.750 0.414 0.510 0.201 0.264
2007.833 0.356 0.550 0.231 0.230
2007.917 0.265 0.490 0.209 0.121
2008.000 0.201 0.400 0.114 0.085
2008.083 0.056 0.130 -0.046 -0.069
2008.167 0.187 0.260 0.020 -0.002
2008.250 0.430 0.600 0.089 0.079
2008.333 0.250 0.410 0.015 0.080
12 mo avg 0.307 0.462 0.148 0.169
Hadley previous
1997.000 0.174
1997.083 0.151
1997.167 0.248
1997.250 0.264
1997.333 0.195
1997.417 0.244
1997.500 0.377
1997.583 0.372
1997.667 0.410
1997.750 0.455
1997.833 0.494 Average
1997.917 0.468 0.321
GISS previous
1997.250 0.46
1997.333 0.34
1997.417 0.32
1997.500 0.50
1997.583 0.26
1997.667 0.37
1997.750 0.41
1997.833 0.50
1997.917 0.56
1998.000 0.53
1998.083 0.52 Average
1998.167 0.79 0.463
UAH previous
1987.750 0.052
1987.833 0.210
1987.917 0.096
1988.000 0.359
1988.083 0.275
1988.167 0.017
1988.250 0.223
1988.333 0.062
1988.417 0.089
1988.500 0.090
1988.583 0.187 Average
1988.667 0.158 0.151
RSS previous
1987.833 0.223
1987.917 0.169
1988.000 0.423
1988.083 0.269
1988.167 0.014
1988.250 0.197
1988.333 0.093
1988.417 0.147
1988.500 0.091
1988.583 0.202
1988.667 0.058 Average
1988.750 0.248 0.178
Here are all four datasets for the past 30 years, with baseline adjustments:
http://www.woodfortrees.org/plot/hadcrut3vgl/last:360/offset:-0.146/mean:12/plot/uah/last:360/mean:12/plot/rss/last:360/mean:12/plot/gistemp/last:360/offset:-0.238/mean:12
This will auto-update as each new month’s data comes out.
(Anthony: If you want easy-to-parse GISS data to answer your first objection, use: http://www.woodfortrees.org/data/gistemp Obviously I can’t answer your second objection!)
REPLY: Thanks Paul, there are days that I wish I was a C++ programmer and could handle these sorts of data nuances. BTW I sent you an email over a week ago. No reponse, you may want to check SPAM filter.
Finally someone has put GISS where it belongs (in the closet) for legal action later lol
… and here’s the latest data, not averaged:
http://www.woodfortrees.org/plot/hadcrut3vgl/last:12/offset:-0.146/plot/uah/last:12/plot/rss/last:12/plot/gistemp/last:12/offset:-0.238
(Incidentally your RSS graph is showing as the zoomed 1998-2008 version, although it’s labelled as 1979-2008)
REPLY: It’s an actual zoom, the graphing program doesn’t automatically change the title when it zooms. The source dataset is in fact 1979-2008
(Oh, that damn 8-followed-by-rpar thing again 8) )
Anthony: I looked at the NCDC Absolute Temperature data sets to see what kind of info they held that wasn’t noticeable in anomaly data. There were many things I hadn’t anticipated, such as the magnitude of the step change in Maximum Annual Land Surface Temperature following the 97/98 El Nino:
http://i28.tinypic.com/11jp09d.jpg
I had expected a subtle change due to the effect of El Ninos on northern high latitudes, but a global maximum LST shift of 0.6 deg C!!!
Due to the number of graphs and the volume of data, I broke it into three posts:
Part 1 – Overview and Combined Global LST & SST
http://bobtisdale.blogspot.com/2008/05/fresh-look-at-ncdc-absolute-temperature.html
Part 2 – Global SST
http://bobtisdale.blogspot.com/2008/05/fresh-look-at-ncdc-absolute-part-ii.html
Part 3 – Global LST
http://bobtisdale.blogspot.com/2008/05/fresh-look-at-ncdc-absolute-part-3.html
I provided TinyPic links under each graph.
I also did a comparison of GISS Temp with 1200 km and 250 km radius smoothing. As you would imagine, the plot of the GISS data with 250 km radius smoothing appears similar to HadCRUT3GL.
http://bobtisdale.blogspot.com/2008/05/giss-temp-smoothing-radius-comparison.html
Regards
Perhaps it has already been treated somewhere, I didn’t found it.
Watching all the graphs in one post like that made me wonder: it seems, roughly, that the trend is flat up to a bit before 1998, then the big step, then flat again.
I don’t know exactly how the anomaly are calcultated, but as far as I understood, at some point there is some average taking into account days/months/years previous and after the point where the anomaly is calculated. That is why times to times some record back in the past may be modified (at least one of the reasons) and why it has been that story recently with the algorithme showing too much warm because the data of tomorrow have to be guessed to calculate the anomaly of today.
If i am not too wrong up to now, then my question is: the step in anomaly observed during the time of Nino 98 could be due just to calculation because the high temperature of 98 are taken into account in the calculation of the post-98 anomaly.
As it should be taken into account in the calculation of the anomaly for the years before-98, my question may not make any sense.
However, if the algorithme to calculate the anoimaly is known (big IF?), does any body calculated what would be the trend of anomaly if the the temperature around 98 peak where taken as zero (for UAH anomaly graph for example, 0.1 for Hadcrut graph and so on). If it has been done (I even don’t know if it is possible…) is there still that step increase or the anomalies after EN98 come back to the level of pre-EN98 level, and the trend becomes flat over 30 years?
Anthony
WRT to the GISS data-set – please note that the +0.67 value for March 2008 released last month has now been reduced to +0.60, the same value as March last year…
REPLY: I took the 2008 data from my previous post on GISS, not the source, not knowing they had revised the data. Will change.
Another reason to wonder what’s up with GISS is that they revised the March anomaly down from .80 to .67. Does anyone know why?
REPLY: Well there was a lot of hollering about GISS March data being such an outlier, perhaps they listened and took another look at the data.
Anthony,
This morning, GISS reports this for 2008:
2008 13 26 60 41
So now March is 60. Is this another revision downward from 67?
your giss is wrong. march 08 has been adjusted down to .6
REPLY: Thanks I took that 2008 data from my last post on GISS, not the source, my bad.
Thanks for the hard work. This is always interesting.
I find it very convenient to compare the four temps on one graph (all using the same reference period). Can you provide that?
Also, while I agree that GISS is an outlier, it is better to see it and compare it to other three.
Anthony, recall that the GISS algorithms rewrite history as future data is added to the pile. As Basil notes above, January was revised upward by 1, and March revised downward by 7.
REPLY: An adjustment here, a revise there, and pretty soon you’re talking about real trends.
While I agree that GISS is an outlier, and also agree that it is a great service to point out its deficiencies, I think it is still worth reviewing. Like it or not, it is the data set that is often looked to as gospel. Know your enemey, and all that. The better we know it the more credibility in critiquing it. Besides, if we can actually demonstrate flat/cooling trends in GISS with all its deficiences and biases, it takes all the argument about data selection out of the mix. In the recent historical contect, and considering the different bases, the differential this month between GISS and other data sets was not remarkably out of line, though it was on the higher-end of reasonable differentials. I do plan on taking a closer look at these with some statistical analysis at some point, though, to get a better idea of what is a reasonable differential and what isn’t.
And Anthony, you disappoint me, my friend. A simple dropping of the data and a parsing of the data – which you need to do even on the single-column data anyways – combined with a vlookup formula is all the one-time setup you need to get the GISS data in the format you desire. I have it all set up in a spreadsheet of my own if you’d like me to send you a copy.
(Since I was just critical, and I know you hate pseudonyms, I’m signing my name at the end of the post… just for you)
Joe Tritz
The Idiot
GISS charts here: http://digitaldiatribes.wordpress.com/2008/05/15/131/
REPLY: Hi Joe, Sorry to disappoint, but for me it’s all about time. I have a national project: http://www.surfacestations.org I’m managing, this blog, a weather technology business I run, a daily radio program, a wife and two young children (that I can’t say “no” to) who often tug me away from the computer at night when I’m writing posts such as this one.
So when I complain I don’t have time to do something, such as to write a program or macro to decode GISS’ unweildy dataset, many people come to the rescue as you did, having already done it. No reason for me to reinvent the wheel. I knew someone would step up and you did. Thanks.
what do all the records look like superimposed on each other? Could you do a post like that soon?
I recall some divergence that Dr. Pielke, Jr. posted.
Sorry if I sounded too critical, Anthony. The comment was intended to be taken a bit tongue-in-cheek (the name of my own blog betrays my sense of humor). I love your blog and all the work you do, and I visit it on a daily basis. No need to explain further.
Just to note for anyone struggling with the unwieldy data formats; anything WoodForTrees can plot is also available as a simple two-column format – just make sure you ignore comments (#…) and the final ‘e’ marker (which divides series if you are doing more than one). You can (as I do) throw this straight into Gnuplot or do whatever else you want with it.
The files are updated at 3am GMT/BST every day from the master sources.
Temperature:
http://www.woodfortrees.org/data/rss
http://www.woodfortrees.org/data/uah
http://www.woodfortrees.org/data/hadcrut3vgl
http://www.woodfortrees.org/data/gistemp
Solar:
http://www.woodfortrees.org/data/sidc-ssn
http://www.woodfortrees.org/data/pmod
CO2:
http://www.woodfortrees.org/data/esrl-co2
PDO:
http://www.woodfortrees.org/data/jisao-pdo
(Sorry Anthony, this is guaranteed to trigger your spam filter, but I thought I’d better put them all in one place!)
Write out 100 times: Must put spaces before right-parenthesis in WordPress comments : )
I see that the WoodForTrees graph offsets both GISS (-0.238) and HADCRUT (-0.146) presumably to bring them into line with the satellite data. I suppose this is really a question for WoodForTrees but maybe it was mentioned somewhere how those offsets were derived? I know how I would compute them but do you know the method?
I am a bit lost. The NCDC publishes a monthly climate report. Here is the month of March:
http://www.ncdc.noaa.gov/oa/climate/research/2008/mar/global.html#temp
If GISS was revised, does someone know if this report would be corrected?
After all NCDC claim that Global Land temperature was the warmest March ever, in their records. It seems that there must be three possible answers.
-The report used correct data to begin with and doesn’t need to be corrected.
-It will be corrected some time in the future.
-They will ignore this information and permanently leave a faulty report.
The answer is C
What’s changed since the end of April? Did GISS suddenly remember they had dozens of unnoticed records, all of them significantly cooler than their extrapolations showed? Did they suddenly notice that they’d read 1000 thermometers too high? Or did they just decide to use a somewhat different algorithm to determine their average?
My best guess would be the last option. But if that’s the case, shouldn’t EVERY average change, from the very beginning of their data? You can’t just pick the algorithm you want, each month. If every data point isn’t calculated the same way, they’re not comparable.
Just as importantly, when do the adjustments stop? At what point can we say, “The high temperature at Denver International airport, on January 7, 1997, WAS 2C?” History is the only absolute in life. It’s the only thing that CAN’T change. If we allow it to, it ceases to become history, and is nothing more than propaganda.
I have a very basic question: The public is obviously being manipulate by GISS, the NCDC, and others on the taxpayers dime. Is it at all possible to file some type of class action suit against these types under the premise of their using taxpayer funds to spread propaganda similar to Hitler?
Just wondering! Any lawyers out there?
Jack Koenig, Editor
The Mysterious Climate Project
http://www.climateclinic.com
rex:
C. is for correct? Or, for corrupt?
I guess we can only watch and wait. I feel a bit helpless against an organization like NOAA, producting these kind of results.
If Anthony ran it, he would have updated that report, with the corrected information, faster than one could say AGW.
D. Quist: Are you asking, will NCDC revise its data to reflect the change in GISS? If that’s what you’re asking, the answer is no. They’re two different entities and they calculate their anomaly data in different manners.
Bob Tisdale
“..NCDC.. calculate their anomaly data in different manner” (than GISS).
Thanks for the clarification!
That leaves two issues.
-How influential is the NCDC report?
-It seems to me, that the NCDC report, that March was the “warmest month ever” is based on some faulty information! If GISS has been adjusted down and the other measurements indicate that March was near average, it would indicate that NCDC is inaccurate.
I have been looking at the NCDC reports for over a year now. I just want to know how accurate they are, and/or how biased their reports are.
Thanks again for helping me understand.
Paul Clark (09:46:11) :
“Write out 100 times: Must put spaces before right-parenthesis in WordPress comments : )”
Or ask Anthony 100 times if Word Press has a mechanism to disable smilies. They’re sort of cute until they corrupt things.
A Preview button would be nice too.
Thank you for not posting 7 seconds later.
Why the changes from GISS?
Missing data filled in.
This April 2008 map is what the Jan-Mar 2008 maps used to looked like.
http://data.giss.nasa.gov/work/gistemp/NMAPS/tmp_GHCN_GISS_HR2SST_1200km_Anom04_2008_2008_1961_1990/GHCN_GISS_HR2SST_1200km_Anom04_2008_2008_1961_1990.gif
Now, look at the Jan-Mar maps and notice the holes that were filled in.
Jan 08
http://data.giss.nasa.gov/work/gistemp/NMAPS/tmp_GHCN_GISS_HR2SST_1200km_Anom01_2008_2008_1961_1990/GHCN_GISS_HR2SST_1200km_Anom01_2008_2008_1961_1990.gif
Feb 08
http://data.giss.nasa.gov/work/gistemp/NMAPS/tmp_GHCN_GISS_HR2SST_1200km_Anom02_2008_2008_1961_1990/GHCN_GISS_HR2SST_1200km_Anom02_2008_2008_1961_1990.gif
Mar 08
http://data.giss.nasa.gov/work/gistemp/NMAPS/tmp_GHCN_GISS_HR2SST_1200km_Anom03_2008_2008_1961_1990/GHCN_GISS_HR2SST_1200km_Anom03_2008_2008_1961_1990.gif
I also am using the HadCRUT base period in these maps. And I would expect more data to be filled in for April and maybe other months in the coming months.
Northern parts of Canada are still missing for Dec07-Feb08, but Mar08 is filled in and seems to be a normal coverage for GISS. (looking back prior to Dec07)
DAV: I posted how I derived these offsets in comments here a couple of weeks ago:
http://wattsupwiththat.wordpress.com/2008/05/05/rss-msu-lt-global-temperature-anomaly-for-april-2008-flat/
(at 10:34:38 on 6 05 08, about 25% the way down)
Given that a lot of people seem to be repeated these figures now I guess I’d better document them on the site!
If you’ve been thinking about this too I’d welcome your thoughts as to whether this makes sense or not.
Paul
wattsupwiththat,
I’m so glad people like you are out plugging away on the horrendous distortion of science that is called “global warming.” I myself have been planning to write a detailed article on the subject, but have been too consumed with the issues that fill my own blog.
I read the book, “Unstoppable Global Warming Every 1500 Years” by S. Fred Singer and Dennis Avery, and by the very first chapter they had already presented a devastating scientific case against anthropocentric GW.
The tactics of those who are using pseudo-science as a means of imposing a radical socialistic redistribution of global wealth is quite frankly shocking.
Ruc ( “Eric”, right? ):
Agreed. A pox on all autosmilies.
Any smilies anyone sees in my posts are actually right-parens!
Emoticons are best expressed via text.
#B^U
I mean “Ric”.
Has anyone look at a plot of the RSS data for all the different channels? ie TLT, TMT TTS and TLS. Plot out the different graphs for Land and Ocean, Land Only and Ocean Only.
My first impression is that there is just a whole lot of interesting stuff to be explained. No one will be anxious to draw a straight line through this data again
Especially the data for TLS. They want to draw a straight line through this data? Seems ridiculous to me.
http://www.remss.com/pub/msu/monthly_time_series/
“Evan Jones (11:49:49) :
> Ric ( “Eric”, right? ):
Yup. There are a few other Rics around, but I haven’t met one yet.
> Emoticons are best expressed via text.
> #B^U
Hmm. I can’t figure out how to hunt that down with Google.
73…
Anthony,
Just some more on the TLS channel of the RSS.
There are two “bumps” on the TLS data. Both coincide with with the two large volcanoes during the satellite data.
The anomaly resets to a lower values than before after the effects of each volcano and then, for all practical purposes, stays constant. There have been no more bumps since Pinatubo. There have been no more volcanoes since Pinatubo. Will the one in South America show up soon in the TLS data? The stratosphere reacts to volcanoes
The satellite TLT data can be interpreted as a period of states from 1979 to 1997, then an unusually strong El Nino, Now having levelled off at a slightly higher temperature. The Lower Troposphere reacts to PDO, AMO etc
These data seem to indicate that changes occur suddenly rather than monotonically (is that the right word) and seem to have little to do with the rise in CO2.
Hmmmmm. Just food for thought.
Last year my local tiny, conservative, yet weird, newspaper got an award for producing an article on global warming and that the sky is falling. Our editor at the time was an out-of-stater woman who was like an orange peel in the middle of a green lawn. Since then we have changed editors who is equally not like us. Nonetheless, he printed a guest column on how global warming has been questioned scientifically, especially regarding global temperature data massaging. I doubt it will get an award for the paper but the letters to the editor have been coming in from all over the world (strange that people from all over the world would read out tiny little hamlet rag) thanking the editor for allowing such a well-written piece.
So in honor of the demise of GW, I would like to propose that we hold an internet wake. To start the roas…er…toast of its demise, I would like to say that GW was such a nice little fellow but a rather gullible. Especially regarding kidney stones. It was later learned scientifically that his kidney stones were actually the result of a tight sphincter.
We are disappointed why so much hocus-pocus science is used in analyzing data. Any averaging carries its method’s bias. We want to determine the CHANGE of temperature taken (1) AT THE SAME DATE AND TIME and (2) AT THE SAME LOCATION. So a meaningful time differential is
T[Montreal May 20, 07 12:00] – T [Montreal May 20, 08 12:00] = Del T/Del (time)
If we had the TIME DIFFERENTIAL FUNCTION for a large number of locations over a 10 year period of time, this function would show reliably what TEMPERATURE changes we really have. The real trends will come form the CHANGE OF CHANGE second time differential.
Are we asking the obvious?
Elementary calculus works in all other fields of science.
No sense is setting one averaging algorithm against another.
All lead to false notions.
Would someone give the source where the RAW location-time-temperature data is available and the rest is simple freshman calculus.
Btw this spring Montreal is 10 C colder then it was a year ago.
So we got one (1) of the time-temp firest diff function.
The four global datasets chart near the top of the page has a shape from about 2002 to now that looks like the output from a square wave generator.
re: temperature difference from one year ago today. I have also wondered about this simple way of depicting change. The live weather stations in Wallowa County show the positive/negative differences from a year ago. Even though last year was cold, the change shown this year, day after day after day, has been in the negative. I second the motion. I want raw real-time temperature readings so that we can generate a graph that shows up or down trends for same time, place, same instrument, but one year later. I would think that monthly averages would be fine for this purpose.
Well guess what. It is 0.1 degrees warmer today compared to last year in Enterprise, Oregon.
REPLY: Unless you are looking at city UHI trends, single stations usually aren’t a good indicator.
Actually I was thinking about all the stations that share common design and placement in each climate “zone” (you know, gardening zones that tell you what plants to have in your garden that can survive the weather). This real, raw data of daily temperature, averaged by month, can then be averaged throughout the zone. The graph would then read that, for example, January averaged 10 degrees above zero in 2007 but averaged 0 degrees one year later. It makes more sense to me that when I am looking at a temperature graph, I can read the actual temperature.