By Steve Goddard
RSS April 14,000 Foot Anomalies
UAH and RSS satellite data have been showing record warm temperatures in 2010, despite the fact that many of us have been freezing – and CAGW types have been quick to jump on this fact. But Had-Crut surface data has 2010 at only #5 through March, which is the latest available. So Watts Up With That?
Had-Crut rankings through March
I don’t have any theories about root cause, but there are some very interesting empirical relationships correlating ENSO with satellite and surface temperatures. Satellite TLT data is measured at 14,000 feet and seems to get exaggerated relative to surface temperatures during El Niño events. Note the particularly large exaggerations during the 1998 and 2010 Niños below. UAH (satellite) is in red and Had-Crut (surface) is in green.
http://www.woodfortrees.org/plot/uah/from:1978/plot/hadcrut3vgl/from:1978/normalise
Looking closer at the 1998 El Niño, it can be seen that both UAH and RSS 14,000 ft. temperatures were highly exaggerated vs. normalised Had-Crut and GISS surface temperatures.
Below is a chart showing the normalised difference (UAH minus Had-Crut) from 1997-1999
The chart below shows the same data as the one above, but also plots ENSO. This one is very interesting in that it shows that the satellite data lags ENSO by several months. In 1998, ENSO was nearly neutral by the time UAH reached it’s peak, and the 1999 La Niña was at full strength before UAH recovered to normal.
Conclusions: ENSO is already headed negative, but it is likely that we will see several more months of high satellite temperatures. CAGW types will abandon their long cherished Had-Crut, and declare 2010 to be the hottest year in the history of the planet. Yet another way to hide the decline.ñ
Bob Tisdale:
I’m sure you are aware that you can find HadCrut temp data here
http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/monthly
but I’m sure similar results should be obtainable with any other global temperature index.
The CO2 data was taken from a combination of the Mauna Loa atmospheric CO2 data and the Law Dome ice core data to extend the data back to the beginning of the HadCrut time series.
Is there any way to post images?
But you don’t even need to go back 160 years. Even from the start of the Mauna Loa dataset in 1958 without the Law Dome extension, the R-sq is 69%.
From 1979 to present, the R-sq is 57%. Then the shorter and shorter a time span you use, the smaller and smaller the R-sq becomes, because the small year to year variability due to CO2 gets lost in the large year to year “natural causes” variability, like El Nino.
However, using a long time period is preferred because it reduces the chance that the correlation is just a coincidence due to CO2 and temp rising at the same time since 1974.
Smokey says:
May 15, 2010 at 8:06 pm
Wren, May 15, 2010 at 6:32 pm,
Hm-mm-m. Who to believe, you? Or Dr Spencer?
There really is no CAGW “hypothesis.” It is simply a conjecture; an opinion, based on always-inaccurate computer models.
To elevate CAGW to the status of a hypothesis, you will need to provide testable, reproducible, empirical evidence quantifying the effect of human CO2 emissions on the planet’s temperature. And a good hypothesis makes accurate predictions. Sorry about your CAGW models, which can’t predict their way out of a paper bag.
So far, all attempts to provide empirical, testable evidence have failed so badly that the proponents of CAGW no longer even make the attempt. Instead, they resort to endless appeals to authority, ad hominem attacks, and the stale old chestnut of “consensus” — a non-scientific argument if there ever was one.
But you, my little chickadee Wren, can be in the history books by providing testable evidence showing the specific fraction of a degree warming per teragram of anthropogenic CO2 emitted. Simple, no?
But if you can’t provide such testable, reproducible measurements, then you are back at the conjecture stage, demanding $trillions based upon your …opinion.
=======
Spenser theorizes fewer clouds cause global warming, If this means the warming trend of the past century was caused by a trend to fewer clouds, I am puzzled. I haven’t noticed that its be getting less and less cloudy where I live over the years.
I’m also puzzled by you presenting a null hypothesis for hypothesis you believe doesn’t exist. You can’t have one without the other. The following is an simplified example of a hypothesis and it’s null.
Hypothesis: Burning fossil fuels has a warming influence on global temperature. Null hypothesis: Burning fossil fuels has no warming influence on global temperature.
Evidence supporting the hypothesis: Burning fossil fuels releases CO2 into the atmosphere, and CO2 is a green house gas that helps keep the earth warm through the green house effect. The null is rejected on the basis of the evidence.
There you have it.
Is that a testable hypothesis? Not in a simple statistical way, as it would be, if for example, my hypothesis was I could taste the difference in Pepsi and Coke at a 95% confidence level in a 20-trial double-blind test, and then was able to reject the null(couldn’t taste a difference) by correctly identify the sodas enough times. However, some hypotheses don’t readily lend themselves to this kind of testing, but do have sufficient supporting evidence to be accepted.
Jbar: You asked, “Is there any way to post images?”
Upload your image to a picture posting site, and provide a link, like this comparison of NINO3.4 SST anomalies versus the monthly change in Mauna Loa CO2 concentration:
http://i33.tinypic.com/2uotpjb.jpg
The graph above is from an old post:
http://bobtisdale.blogspot.com/2008/09/atmospheric-co2-concentration-versus.html
The reason I asked about the correlation is that I would have expected it to be higher. I’ve correlated CO2 and monthly global temperature in the past and found the monthly data to have a higher correlation coefficient than what you’ve listed. And as illustrated above, the correlation is a function of CO2 Concentration’s dependence on sea surface temperature, not the other way around.
Wren says:
May 15, 2010 at 10:50 pm
Evidence supporting the hypothesis: Burning fossil fuels releases CO2 into the atmosphere, and CO2 is a green house gas that helps keep the earth warm through the green house effect.
Is that a testable hypothesis?
Yes it is testable, it has been tested using advanced mathematical statistics by a lot of Scientists and mathematicians and found to be completely false. There is no Correlation, just have a look at Bart’s thread here
http://ourchangingclimate.wordpress.com/2010/03/01/global-average-temperature-increase-giss-hadcru-and-ncdc-compared/
Read the work by VS and Tim Curtin which show now correlation either on the Global or Local scale.
But of course you STILL won’t beleive it, will you?
Joel Shore says:
May 15, 2010 at 4:07 pm
In much of the middle and southern part of the U.S., they did experience colder than normal temperatures this winter, but places in Canada were above normal.
Being an American, you seem to think the Northern Hemisphere is the US Continent, you seem to have forgotten the whole of Europe, Russia, China and India, all of which had record breaking Cold weather.
You also say
However, temperature anomalies show a significant positive correlation even out to distances of about 1000 km. Hence, if one uses temperature anomalies, one is able to get good results for an average with much fewer, widely-spaced measurement locations than one can with temperature.
Yes but anomalies to what? The whole dataset or a selected time span?
So you are saying that the less datapoints that you use the more Accurate the results?
This Statement “The analysis method was documented in Hansen and Lebedeff (1987), showing that the correlation of temperature change was reasonably strong for stations separated by up to 1200 km, especially at middle and high latitudes.”
Reasonably Strong does not cut it in Statistcs.
Somebody else did what I did combining Lawdome and Mauna Loa vs. Hadcrut. Scatter plot here:
http://chartsgraphs.files.wordpress.com/2009/08/co2_temp_scatter_regression.png
And the post in which he discusses it here
http://chartsgraphs.wordpress.com/2009/08/13/excel-chart-misrepresents-co2-temperature-relationship/
The “misrepresenting chart” to which he refers is not the scatterplot above, but a short-term (1995-2009) comparison plot by a “Senator Fielding” with the CO2 scale exaggerated WELL beyond its actual relationship with temperature.
Incidentally if you do a scatterplot of the annual average data instead of monthly data, the R-sq goes up to 77%, but really only because the monthly short term variability is taken out.
Bob Tisdale:
In your May 16 2:07 AM you link to a comparison of CO2 rate of change with various temperature data, including Nino 3.4 SST anomalies. I’m not familiar with the Nino 3.4 data series(es). Doesn’t it represent the temperature DIFFERENCE between eastern and western areas of the tropical Pacific and not the actual sea surface temperature?
I have not searched the exact cause of the seasonal variability in CO2, but obviously it is associated with the time of year.
Jbar: May 16; 5:12 am
The forcing impact of Co2 is logarithmic with respect to its level and then the resulting Forcing itself is also logarithmic with respect to its impact on temperature.
F=5.35 ln(CO2 now / CO2 orig)
Temp = ((F orig + F new) / 5.67e-8)^0.25
A linear function works okay right now because we are in a particular section of the function that is something close to linear, but it starts to move off for CO2 levels lower and higher than the range used.
A C Osburn says:
I was using the North American continent as an example of some people looking very selectively. What is your evidence that “the whole of Europe, Russia, China and India, all of which had record breaking Cold weather” and why is it that neither the surface temperature record nor the satellite record seems to support your claim for the Northern hemisphere as a whole?
The baseline period for the anomalies essentially just represents a constant offset. Why does it matter what you choose for that baseline? And, no I am not saying that fewer data points give more accurate results. What I am saying is that, given the limitations on the amount of data that you have, using the data in an intelligent way can give more accurate results than using it in a less intelligent way.
Why don’t you read the original paper then? Arguing that a one-sentence summary of a scientific paper does not have enough detail is a pretty lame claim.
What a coincidence, I found someone on the web arguing pretty much the very same thing that Smokey is:
Okay, I admit that I had to get a little creative regarding the part in brackets. The actual quote can be found here: http://www.gty.org/Resources/Articles/A176_How-Important-Is-Genesis-13 The fact that these sorts of arguments can be made against science that I think most of us agree is good science shows how weak such arguments are, particularly when they contradict the conclusions that just about all of the major scientific academies and scientific societies have reached.
Bob Tisdale – you said
“BTW: If you assume the oceans integrate the effects of ENSO, there is no need to introduce anthropogenic greenhouse gases to explain the rise in global temperature.”
The issue is that in a multiple linear regression of global temperature and climate variables, CO2 is the most dominant factor. Then there is the fact that CO2 does absorb infrared, giving it a compelling physical reason to be correlated with temperature. Maybe a model can be constructed without it, but you’re not supposed to throw out data, not even one point, without a darn good reason, let alone the biggest model factor.
I’m saying this without having read your link yet. (Time, time, time. Retirement is not that close.)
A lot of strange ideas in here now. The derivative of SST almost equals -SOI (minus volcanoes) i.e El Nino heats the global ocean and SST is the integral of -SOI.
http://virakkraft.com/SOI-SST-deriv.png
*******
John Finn says:
May 15, 2010 at 3:57 am
I think both you and PearlandAggie need to get the area size of urban developments relative to the rest of the world into perspective. You could drop the top 20 largest cities in the world into Texas and not notice they were there. The reason that UH *might* have an impact on surface temperature trends is because, for obvious reasons, the stations happen to be close to urban areas.
*******
I’d have to agree w/John Finn on this. The spatial coverage of urbanized areas is very small, certainly less than 1% of the globe. However, for very densely urbanized regional areas like much of Europe, eastern China, Indonesia, the US megalopolis from DC to Boston, etc., the UHI effect might be picked up by satellites when the air is rising from these areas. But the total effect globally would be quite small, I’d think.
Bob Tisdale, in RE my May 15, 6:29 AM “it’s pointless to look at month to month temperature data, or even year to year, and then argue about whether anthro GH gases are responsible or not” “We have to wait DECADES to see if there’s been a real change [in the warming rate]”
My plot which I had in mind on this comment conveniently shows up in A C Osborn’s 5/16, 3:59 AM link:
http://ourchangingclimate.files.wordpress.com/2010/03/global_temp_yearly_p15_trendline_tavg_incl_txt_eng.png
The three diagonal rising lines on the right are the trendline from 1974 to present, and the upper and lower “control limit” on the data. (My lines are a little different because I used HadCrut only, but it’s the same concept.)
The linked plot is not updated to the present. Right now global temp is right on the centerline.
Every process has natural variability. This kind of method called “control charting” is used to monitor that variability. It’s mainly only when the width or value of the range changes that you know for certain that something has changed in your “process”.
The uptrend is about 0.2 deg C/decade. Not all of this is attributable to CO2, by my regression only about 0.13/decade is.
In this linked plot, this fellow Bart only shows +/- 0.2C variation above and below the trendline, where as my analysis gives a range of +/- 0.4C. This is because he is using annual data here whereas my plot (not linked) is to monthly data. If you compare the up/down gyrations of the plot (especially the monthly data) with El Nino indicies, it’s visually clear that El Nino seems to be driving much of this +/- 0.4C variability. (El Nino gets very spiky!)
So my point was, year to year temp varies over a 0.8C window, thanks to natural variation. As long as the “process” doesn’t change, it should stay that way. In fact, it is pretty consistent even if you go back in time to 1850. The temperature range is typically about that much above and below the moving-average trendline even way back then.
However, the yearly change in the trend line is only one 40th of that range, or +0.02C per year. It takes 20 years for the centerline to move half of that range. So you can’t possibly tell if the trend is changing direction from one month or one year of data. Even if the rising temperature trend comes to a dead stop, there is so much intra-year and year to year variation that you won’t even be able to see it for 5-10 years at a minimum.
Just look what happened in the recent past for proof of this. Take 1992 for an example. Temperature dropped really low that year and it looks on this plot like it was curtains for global warming. (Of course, this was probably Pinatubo related but…) Then over the next 5 years, instead of GW ending, it shot up to its 1998 record high.
But did it keep going up? No, it moved back down to the trendline and in winter 2007/8 it got smacked down to the bottom of the range again due to a brisk La Nina. Did it keep going down in 2008 and 2009? No, it rebounded back up to the trendline again where it is today. It’s doing what it normally does, bouncing around A LOT.
As long as global temperature is bouncing around inside this rising trend range, around the centerline, the rising trend is still “intact”. To really be sure the trend has changed, the data should start breaking out above or below the “control limits” for a sustained period of time (i.e. not just briefly like it did in hot 1998 and cold 2007/8 winter). Recent data however still has temperature bouncing around that centerline, THEREFORE THE RISING TEMPERATURE TREND IS STILL INTACT.
It may look to our eyeballs like the trend has flattened out since 2000, but it has done things like this before since the trend began in 1974 AND RECOVERED to keep on rising. There is SO MUCH short term variation in this data, that it is not statistically possible to falsify the null hypothesis that “global temperatures are rising at 0.2C per decade”. That will take AT LEAST another 5 to 10 years of data. If the trend rate changes only a little bit, say from 0.2C per decade to 0.1C/decade, it may take even longer than 10 years to declare that warming has officially slowed.
DON’T GET ME WRONG!
I watch the monthly temp and ice coverage data just like a lot of you guys probably do. (I check the quarterly new case rates in the global war on polio, too. And Spaceweather.com for NEO close encounters and the sunspot of the day.) It’s like a sport, if a tediously slow one, and you cheer for your own team when it goes your way and get a little glum when it doesn’t. BUT you have to remember that you can’t rest assured that you’ve “won or lost” based on just one month or year of data.
P.S.: Is there a different way than CAPS to show emphasis? I don’t mean to shout, I just want to provide some inflection.
A C Osborn says:
May 16, 2010 at 3:55 am
Wren says:
May 15, 2010 at 10:50 pm
Evidence supporting the hypothesis: Burning fossil fuels releases CO2 into the atmosphere, and CO2 is a green house gas that helps keep the earth warm through the green house effect.
Is that a testable hypothesis?
Yes it is testable, it has been tested using advanced mathematical statistics by a lot of Scientists and mathematicians and found to be completely false. There is no Correlation, just have a look at Bart’s thread here
http://ourchangingclimate.wordpress.com/2010/03/01/global-average-temperature-increase-giss-hadcru-and-ncdc-compared/
Read the work by VS and Tim Curtin which show now correlation either on the Global or Local scale.
But of course you STILL won’t beleive it, will you?
=====
I doubt Bart, VS, or Curtain claim rising atmospheric CO2 levels don’t have a warming effect on global temperature. If any of them do, please quote them.
lgl says:
May 16, 2010 at 6:39 am
A lot of strange ideas in here now. The derivative of SST almost equals -SOI (minus volcanoes) i.e El Nino heats the global ocean and SST is the integral of -SOI.
http://virakkraft.com/SOI-SST-deriv.png
_______________________________________________________________________
Can you flip that transparency so the writing is not mirror image? Or at least reproduce the legend so the lines make sense to those trying to read the graph. And yes some of us do look at the references.
Joel Shore is so inept that he actually believes posting his religious tract on the “Best Science” site is a valid argument. It’s not. I could post a tract on Scientology and say, “that’s Joel.” Come to think of it…
As usual, Joel Shore’s psychological projection causes him to assign his own faults onto others. The True Believers in catastrophic AGW, like Joel Shore, base their belief in approaching thermogeddon on blind faith, simply because they must: faith is all they’ve got, because CAGW is not supported by any testable evidence, or by validated models capable of making accurate predictions, or by the planet itself. It is simply a conjecture.
When Joel says “most of us agree,” he must be talking about him and the mouse in his pocket. Because his constant appeals to authority are all he’s got; testable facts supporting CAGW are non-existent, and consensus is an illogical argument, only employed because there is no empirical evidence showing that humans can control the climate. But you play the cards you were dealt, and CAGW is a fact-free conjecture. Time to fold.
Gail Combs,
Sorry, the original is here: http://www.cgd.ucar.edu/cas/catalog/climind/soi.html
SOI=-1 gives roughly 0,1 C/yr SST increase. (the woodfortrees SST graph is monthly)
davidmhoffer says:
May 15, 2010 at 12:08 am
“So I think I get it.
Warming causes increased convection so warm air goes up, mostly in the warmest zones and pulls cold air in behind it mostly from the colder zones. So warm air goes up at the equator and spills over sideways somewhere around 14000 feet or so, so they can lie on their backs waving at the satellites and screaming hey look at us! In the meantime air from the arctic zones has to come in underneath to replace the warm convecting air, and so cools off the land surface as it goes. Ergo the earth is warming up and the nasty cold summers for the last couple of years are evidence.
So how hot does the planet have to get before it starts snowing in Florida in July?”
_______________________________________________________________________
You forgot to mention that as the warm air goes up (think Cumulonimbus ) the vapor condenses to rain and puts all that heat that got sucked up from the ocean into the atmosphere closer to escaping to outer space.
Somehow the emphasis is always on the heat minor percentage of heat radiated back down from that piddling little .03% CO2 and not on the heat rapidly transported up by thunderhead formation.
And yes I know you were being sarcastic.
Wren says:
May 16, 2010 at 7:40 am
I doubt Bart, VS, or Curtain claim rising atmospheric CO2 levels don’t have a warming effect on global temperature. If any of them do, please quote them.
Wren, I pointed you to the truth, if you are too lazy to read it foryourself I am not going to do it for you.
Both VS & Curtin (NOT Curtain) don’t claim, they PROVE no correlation.
Joel Shore says:
May 16, 2010 at 5:45 amWhat is your evidence that “the whole of Europe, Russia, China and India, all of which had record breaking Cold weather” and why is it that neither the surface temperature record nor the satellite record seems to support your claim for the Northern hemisphere as a whole?
Exactly my point, as humans we experinced the cold, it was Headline news, we have our own thermometers to take measurements with to confirm, seasonal records broken in the US, Europe, Russia, Mongolia, China. So just why is the Official Temperature Set and the Satellite temps so at odds with REALITY.
Jbar: You replied. “Then there is the fact that CO2 does absorb infrared, giving it a compelling physical reason to be correlated with temperature.”
The problem: The planet’s surface is 70% ocean and only the top few millimeters of the oceans absorb Downward Longwave (infrared) Radiation . So the impact of any additional DLR due to anthropogenic greenhouse gases will be released as evaporation.
Jbar: You wrote, “I’m not familiar with the Nino 3.4 data series(es).”
The NINO3.4 region is located in the central equatorial Pacific, coordinates 5S-5N, 170W-120W. NINO3.4 SST anomalies are one of the primary measures of the sea surface temperature component of ENSO. It is the basis for the NOAA ONI Index, which is a 3-month running-average of NINO3.4 SST anomalies.
John Finn says: May 15, 2010 at 3:57 am
I think both you and PearlandAggie need to get the area size of urban developments relative to the rest of the world into perspective. You could drop the top 20 largest cities in the world into Texas and not notice they were there. The reason that UH *might* have an impact on surface temperature trends is because, for obvious reasons, the stations happen to be close to urban areas.
beng says:May 16, 2010 at 6:49 am
I’d have to agree w/John Finn on this. The spatial coverage of urbanized areas is very small, certainly less than 1% of the globe. However, for very densely urbanized regional areas like much of Europe, eastern China, Indonesia, the US megalopolis from DC to Boston, etc., the UHI effect might be picked up by satellites when the air is rising from these areas. But the total effect globally would be quite small, I’d think.
What does spatial coverage have to do with anything? It’s the magnitude of the warming, unrelated to “area”. Considering a “well mixed” atsmosphere, does a 1 km\2 erupting volcano or a 1 km\2 coal plant put out the same thermal energy as a 1 km\2 parking lot?
A C Osborn says:
May 16, 2010 at 8:46 am
Wren says:
May 16, 2010 at 7:40 am
I doubt Bart, VS, or Curtain claim rising atmospheric CO2 levels don’t have a warming effect on global temperature. If any of them do, please quote them.
Wren, I pointed you to the truth, if you are too lazy to read it foryourself I am not going to do it for you.
Both VS & Curtin (NOT Curtain) don’t claim, they PROVE no correlation.
=======
That’s debatable, but beside the point anyway. The issue is causality not correlation.
You brought up Bart, VS, and Curtin. If they think their analyses proves CO2 from man’s activities does not have a warming affect on global climate, I presume they say it, and will not mind you quoting them here. So please let me know.