I’ve mentioned before how chosen color schemes greatly influence how people see surface temperature data. Frank points out that sea surface temperature presentations suffer from the same problem. – Anthony
Guest post by Frank Lansner
This is no news – but still needs to be told. NOAA can in many contexts come up with the hottest temperatures available. Here we take a look at the European Sea Surface Temperatures as of 3 may 2010.
NOAA vs. UNISYS, SST, Europe. When I look at this compare, again and again I have to check if these SST are from the very same date, 3 may 2010. But they are. Differences are immense to an extend where it hardly makes sense to look after the European SST?
NOAA is hotter than UNISYS in for example these waters:
The Baltic Sea, The North Sea, The Caspian Sea,
And in addition,
The Black Sea has NOAA Approx. 3,5 K warmer than UNISYS, and
“The NOAA hotspot area” North of Scandinavia: NOAA Approx. 4 – 6 K warmer than UNISYS .
Is there a valid sound simple explanation for these great differences?
In addition NOAA uses a colour scheme that makes Europe look as if surrounded by burning lava. It’s quite a difference to the impression you get when looking at the UNISYS graphic.
So which graphic is correct? For the Baltic, here’s what the “jury” says, SMHI (From Sweden) has an updated SST for the Baltic Sea from exactly 3 may 2010:
The 3 graphics agree reasonably for the Northern Baltic Sea, but for the rest of the Baltic Sea, SMHI shows in average around – 1,5 degrees Celsius anomaly. Both UNISYS and NOAA show too warm temperatures, but NOAA far worse than UNISYS. So, NOAA is around 2 K warmer in this area than SMHI – the best estimate.
Europe is not the only area where NOAA has warmer temperatures than UNISYS. NOAA appears markedly warmer than UNISYS on the Northern Hemisphere – but a little colder than UNISYS in areas of the Southern Hemisphere:
http://weather.unisys.com/surface/sst_anom.html
Link to NOAA SST – use “FULL GLOBAL” to see all:
http://www.osdpd.noaa.gov/PSB/EPS/SST/climo.html
Link to SMHI detailed SST for Baltic + Danish waters:
http://www.smhi.se/polarview/
============================
Visit Frank’s blog: Hide the decline
Frank,
Half of you all have dropped the hint – but let me spell this out:
The baselines appear to be different. It is not enough to say “anomaly” but “anomaly from 19xx to 19yy average” as a baseline.
Other differences – like night vs. day time measurements, have also to be looked at.
On the proper color I have to agree with RobJM as this would help tremendously identifying the maps as anomalies and not temp.
This just highlights the whole sad anomaly fiasco.
The way it is now everybody can have their own movable baseline.
This is what makes the anomaly so confusing and deceptive.
I say go strictly with K units or use some universally agreed standard temperature and pressure like 25 C and 760 mmHg etc.
It is like changing par on your golf score.
If you normally shoot in the 90s you could say “Hey I’m only +15 today.”
The answer is “No you slug you were 33 over par meaning you shot a 105. So shut up and quit trying to embarrass yourself and everybody around you.”
Could NOAA’s red be the new blue?
“we understand differing baselines just fine, the issue is the color presentation. -A”
Reading the article, the author more than once directly compares anomaly readings from the different maps. So this article is not just about color presentation. It is saying the data sets are actually inconsistent with each other.
The word ‘baseline’ doesn’t appear once in the article.
So for the readers who aren’t familiar with these particular data, could you tell us what the baselines are for the three different maps?
For SMHI I see this, “15-year reanalysis using HIROMB for the period 1990-2004”
On a quick look, I couldn’t find the baseline for the other two.
I agree. Use absolute temps and all should agree to use a standard color scheme. Afterall, THEY are asking US to do far more agreeing to do this or that. Seems a small price to pay to win over our cooperation.
Anthony/Frank,
You guys are getting more desperate by the minute complaining about the colour schemes on the NOAA maps. In any event, the NOAA colour scheme is far easier to make sense of to the non-expert with yellow to red showing warmer and blue to purple showing cooler. Since when has green or even blue ever represented warmer than a baseline? Does the hot tap in your bathroom have green or blue above it??????? Does a children’s temperature thermomoter show warm as green or blue? of course not. As a non-expert in this field I cannot make any sense of the UNISYS colour scheme.
REPLY: Yeah sure, whatever.
Yes, without a common baseline, anomalies become almost meaningless — especially if one is interested in the science of changing conditions (I fear to say the dreaded phrase “climate change”) and comparing one method of measurement to another. Perhaps it is time to do away “anomalies”? Are they really more useful than actual temperatures if anyone can cherry-pick? Of course, there is outright lying re the “actual temperatures”. NOAA anyone?
I wish Leif would apologize to Frank. The former is usually exact in relationship to the current science, so many of us notice his posts in particular. And Frank Lansner has consistently offered some of the finest posts on WUWT for my money.
Both NOAA and UNISYS measure anomalies. I check UNISYS daily; the chart is both informative and absolutely beautiful. A morning feast for the eyes.
Tom in Florida says:
“Why do different groups use different baselines?”
If your dataset starts more recently than somebody else’s, then you can’t use the baseline used by the older datasets. Meaning, when RSS and UAH first started publishing their numbers, they simply couldn’t use the baselines normally used by GISS or CRU, because there are no satellite data for 1950. On the other hand, should we expect GISS and CRU to start using a new baseline every single time a new satellite record comes into being? This would also cause confusion, as then GISS publications from 1990, 2000, 2020 would be on different baselines from each other.
So it’s just best to take note of what baseline any given dataset is using, and if you want to make things directly comparable, add or subtract the necessary offset, before continuing. That’s easy enough to do.
In any event, the article above is very remiss for not telling the reader what the baselines are.
Hot is hot – I don’t think the color schemes really matter:
http://seattlesportsnet.files.wordpress.com/2009/04/marisa_miller_surfing.jpg
http://i81.photobucket.com/albums/j215/xiaoroy/Bikini/HotBikiniGirl-YuzukiAikawaandChi-19.jpg
http://www.dailystoke.com/wp-content/uploads/2009/02/bar_rafaeli1.jpg
Carrot Eater:
“So for the readers who aren’t familiar with these particular data, could you tell us what the baselines are for the three different maps?
For SMHI I see this, “15-year reanalysis using HIROMB for the period 1990-2004″
On a quick look, I couldn’t find the baseline for the other two.”
My point exactly. What are they comparing these temps to.. winter of 1974 or 1800 when the Little Ice Age ended? In both cases, of course it’s warmer. But to not include a reference to their baselines doesn’t make their maps any more scientific than musing over the taste of your morning coffee..
“husten says:
May 7, 2010 at 5:58 am
Frank,
Half of you all have dropped the hint – but let me spell this out:
The baselines appear to be different. It is not enough to say “anomaly” but “anomaly from 19xx to 19yy average” as a baseline.”
Husten, this is 100% correct, and thankyou for writing. But if you look at the compare for the whole world, the 3´rd graphic above, you will see, that on the southern hemisphere, NOAA is in fact a little colder than UNISYS as I write in tha article. And we cant have a baseline doing one thing in the NH and another in the SH, see?
Please take a close look:
http://hidethedecline.eu/media/SSTNOAAvsUNISYS/SST2010may3World.jpg
The overall differences between NOAA and UNISYS are not thaaat big in the overall picture, but here are features specific places on the globe where NOAA just differ wildly from UNISYS within specific regions. Take a look at the central North Pacific, the blue oval.
Around the blue oval the NOAA and UNISYS are pretty much shoing the same, but IN the oval, things goes bananas. These issues are clearly NOT baseline issues .
And the examples i have given for Europe where sea areas are 1,2,3,4,5,6 K warmer than UNISYS cannot be baseline issues either.
Just ONE glimpse at the NOAA North Scandinavian hotspot – that has been there for years – tells you that there are some data issues that really needs attention, and that has nothing to do with small differences in baselines nor colore schemes etc.
The Black sea, NOAA´s Black sea is over 3 K warmer than UNISYS. If such a thing was a baseline issue, i would very very much like to know what years their baseline was taken from 🙂 – perhaps year 9000 b.c. – 8900 -b.c. ?
K.R. Frank
Don’t we have a STANAG to cover this type of confusion so that we use the same colors to represent the same data? I thought everything and anything was covered by a STANAG. The Chinese didn’t call us a Paper Tiger for nothing.
Slightly OT
Pielke Sr has a post on cooling ocean tropical temperatures and the prospects for cooler ENSO in the fall.
http://pielkeclimatesci.wordpress.com/2010/05/07/recent-variations-in-upper-ocean-heat-content-information-from-phil-klotzbach/
No credible baseline can explain the NOAA map. For example the sea south of Åland is shown as being warmer than normal and the Gulf of Finland as being colder than normal. The actual water temperature was about 3 degrees centigrade in both areas on May 3 (http://www.smhi.se/oceanografi/istjanst/produkter/arkiv/sstchart/sstchart_20100503.pdf).
This means that the NOAA baseline must have warmer temperatures in the Gulf of Finland than in the Northern Baltic, a situation that simply does not occur in spring.
Incidentally differences between daytime and nighttime water temperatures in the Baltic is negligible this time of year.
There are substantial differences in the reported anomalies for areas like the Black Sea, Faroe Is., and N. of upper Norway. Either NOAA and UNISYS are using different data, or else they are using dramatically different reference periods to compute their anomalies. Some commenters have suggested that the latter is the case, but they have provided no actual information about the nature of the differences.
As for the color schemes, I don’t like either. NOAA reasonably uses yellow for neutral (or just above neutral), but then makes no use of green, and uses black for both -4 and land. Green would be a good alternative color for neutral. INSYS has even green way up at 1.75, and so fails to exploit the red region. The shades of magenta are also hard to distinguish.
Perhaps NOAA is trying to accommodate common RG colorblindness (10% of males?) by not using green. Even so, there must be a better way to do the blue end of the color scheme. Perhaps there is some way brightness could be used to distinguish greens from reds without giving up on green altogether.
But if yellow = R+G, does this register as R+R if you’re RG colorblind?
Using light gray for land rather than black (which is also in the color scale) would be a help.
The NSDIS baseline seems to be based on the dataset here:
http://www.osdpd.noaa.gov/ml/ocean/sst/monthly_mean_sst.html
That means it comprises 1984-1998, except for a hiatus june 1991 – march 1992, which was deleted because of Mount Pinatubo cooling. Looking at the data suggests an explanation for the weird permanent “hot-spot” north of Norway. For some reason this area is missing in most of the monthly maps, so presumably in this are the anomaly is calculated on a much smaller dataset, or simply interpolated.
The NOAA baseline must be after 1972 or 1979, since this is the starting point of the measurements, i believe. But in the decription I have not found it, perhaps missed it:
http://coralreefwatch.noaa.gov/satellite/methodology/methodology.html#ssta
This means, that NOAA has rather recent baseline years, hot baseline years, and baseline is not helping to explain the very hot specific areas here and there on the North Hemisphere.
And as i said, baseline differences (especially when NOAA uses warmer years) can only explain a fraction of these odd warm features.
The pearl above all is the NOAA hotspot north of Scandinavia up to 6 K over UNISYS, any baseline issues could only explain a few % hereof.
“And as i said, baseline differences (especially when NOAA uses warmer years) can only explain a fraction of these odd warm features.”
Since you wrote the article, it’d be helpful if you actually completed the circle and chased down what the baselines actually are, so that we can compare apples to apples, instead of guessing about it.
If any difference remains, then we move on to what the data sources are. Are they using different sources, or is it the same source with different processing? Is this all satellite data? Is there any model re-analysis in there?
Anomaly data for a specific day is in itself interesting. Is the baseline all May 3rds within the baseline? All May days within the baseline?
Nick Stokes says:
May 6, 2010 at 11:45 pm
The color scheme fussing is silly. My own area is computational fluid dynamics, and the NOAA style scheme has been standard for at least 30 years. It’s even there as one of the standard palettes (heat.colors) offered by R, a stats package which has nothing to do with climate science.
Yes, it makes higher temperatures look hotter. It’s meant to. It communicates something to the reader.
And Leif and Rattus are right. If you’re doing these comparisons you should at least acknowledge the difference in anomaly bases.
All beside the point. Well, maybe not. To reiterate what you said:
“It communicates something to the reader.”
Yes, it does. It communicates, to use the phrase in the original article, you are all going to die in a burning pool of larva.
It’s akin to the little brother holding his snot covered finger 1 inch from his sisters face and calmly repeating “I’m not touching you; I’m not touching you; I’m not touching you…”.
All too often the scientific comeback is that it’s not their fault you (the general public) don’t understand. Actually, yes, it is the scientists fault.
But it goes beyond ‘fault’ – it comes down to complete lies. A lie of omission is still a lie and is very easy for scientists to do, deliberately or otherwise, because the audience is not a specialist in that field; they cannot ask the correct questions to determine if an omission has been made. Luckily, science is science. Those whose scientific background is in another field which relies on the same techniques can intelligently ask those questions because they can identify flaws in reasoning, even if they may not understand fully the subject matter. This is where the (climate) ‘scientific’ comeback is – “well, you aren’t a climate scientist…”.
“carrot eater says:
[…]
If your dataset starts more recently than somebody else’s, then you can’t use the baseline used by the older datasets. ”
Why not?
Hi Carrot eater!
Promise me one thing: Its fine and good that you want all relevant questions answered, but: Please don’t overlook 6 K difference in anomaly because you focus on baseline issues that might explain 0,1K ok? Promise? Because then my bringing these fatal data issues to you guys will fall on the floor, and NOAA should not get away with this so easy.
And then one more time: I can’t find the baseline years.
But if you then say: “FINE, no baseline years written by NOAA, then lets forget a 6 K anomaly North of Scandinavia!” i will be surprised.
The satellite data from used by NOAA and UNISYS are likely from the satellite years and thus comparable. And don’t forget that NOAA’s world map is in fact older in the Southan hemisphere than UNISYS as i write above in the comments.
Therefore, the baseline cannot be that different between UNISYS and NOAA.
What we look at here is Freak data limited to specific areas. These freak-warm specific areas cannot be explained by a baseline issue, even though it would be good to know the baselines.
correction, went too fast:
“And don’t forget that NOAA’s world map is in fact colder in the Southern hemisphere than UNISYS as i wrote above in the comments.
Therefore, the baseline cannot be that different between UNISYS and NOAA.”
Sorry if this is Off Topic, but:
It occurs to me that agricultural land usage could be a useful calibrator of these allegedly foolproof satellite-measured time series. (After all, what greater contrast could there be between the overzealous climatologist who never goes outdoors, and a pragmatic no-nonsense farmer.)
Is anybody aware of data showing “crop range” – the lands where farmers think they get optimum yields over the years? Such data would of course be subject to other influences such as market price, new strains, etc., but intelligently treated and filtered might contain a useful climate signal.
Anu says:
May 7, 2010 at 6:42 am [ … ],
For once we agree.
Frank,
“Please don’t overlook 6 K difference in anomaly because you focus on baseline issues that might explain 0,1K ok?”
Until somebody (hopefully you, since you’re the one who raised the point) actually goes and puts the two datasets on the same baseline, none of us know whether the difference due to different baselines is 0.1 K or 3 K or what. Keep in mind that the difference in baselines will be a function of both location and probably also day, if not week or something like that.
So until you do that legwork, it’ll be difficult to know what we’re looking at, if anything. If these were the more familiar land station datasets, somebody else would have done it by now. But these are data I’m not familiar with, and I doubt I’m the only one.
Hand-waving about the baselines is not sufficient here. Please first do the math. Then we’ll see what difference is left, and why that might be.