The Blackboard

Thanks Lucia,

The point of the post was to investigate whether differences between UAH and RSS were significant after June 2002. The reason for looking at this timeframe is due to the use of the UAH station keeping satellite which doesn’t require the massive corrections to trend created by diurnal drift. I currently believe UAH since 2002 is the only reliable trend measure we’ve ever had.

Chad at treesfortheforest uncovered the fact that the diurnal corrections were causing huge discrepancies between UAH and RSS over land – like 2 C/decade since 2002. The fact that the new UAH AQUA sat is for the first time not requiring corrections is substantial. Here are 3 links left at WUWT which describe and answer some of the problems with what in this case is RSS data.

http://noconsensus.wordpress.c.....e-metrics/

Which was demonstrated first by Chad at treesfortheforest and was followed by this set of replies by Dr. Christy:

http://noconsensus.wordpress.c.....ng-closer/

And is the subject of a recent paper presented at Roger Pielke’s blog here:
http://pielkeclimatesci.wordpr.....f-arizona/

If you read the links you find that RSS and UAH used a balance of day and night satellites to keep diurnal correction problems in check in the past. Currently RSS only has NOAA-15 to work with and no counterbalance while UAH (which used to use the exact same satellites as RSS) has begun using AQUA that crosses the equator at the same time every day.

Ok, so the point is that when I ran across this significance, it was purely by accident. One difference though between what you have done and what I did was to re-anomalize the data after it was truncated. It was done for the differencing process after 2002 and left in for the RSS trends. You’ll find a slight increase in significance from that I think.

In the end, it means RSS has a problem which has existed since inception. Changing the corrections for NOAA15 will require recomputing all the historic signal as well which may even affect the 1992 transition which favors the UAH trend also in my opinion. Dr. Christy did a good job explaining what’s happening.

JeffID–

Yes. I thought your point was the difference between RSS and other observations. But Anthony’s title focuses on the observation of the cooling trend. People do read titles and I wanted to look at the issue prompted by that title.

I have always preferred the UAH methodology so I’m pretty skeptical about the recent negative drift in RSS. Jeff has had email conversations with John Christy who reckons that RSS is over correcting for diurnal drift and that, currently, that is yielding too much cooling, since they are presently using NOAA-15 only. UAH is using AQUA, so the station keeping thrusters make diurnal drift negligible.

http://noconsensus.wordpress.c.....ng-closer/

I interpreted his title as kind of a poke at Tamino and RC who have said several times, there is no cooling because it’s not significant.

From my recollection, there were several years which showed significance after re-anomalization but none were very strong. The difference between UAH and RSS after 2000 is just an actual problem with RSS IMHO though.

Jeff Id–
Ahhhh!!! Well, the mumblings that the negative trends aren’t “real” because they aren’t statistically significant is what makes seeing even this sort of ‘interesting’. Maybe not scientifically interesting, but interesting from the point of view of rhetoric.

As a rhetorical tidbit, this will be a talking point until El Nino increases or persists long enough to make the ’significance’ go away or until RSS decides there is something wrong with their processing method.

Well, it the past, we do seem to have read some weird criteria for assessing whether or not a trend is meaningful. I’m under the impression ’some’ have sometimes suggested we can’t reject 2C/century unless we have enough data to reject 0C/century.

If this criteria were true, then, generally speaking we would be unable to reject the hypothesis that the trend is 2C/century when the real climate trend is honestly, truly and absolutely 0C/century!

Sorry to nitpick, but where you wrote “The October anomaly of 0.0282C” in the first paragraph, do you mean “0.282″?

Doug– Thanks!

What makes this REALLY fun for us Deniers is that so many have hammered us that RSS is better than UAH. OK, there is significant cooling by their preferred index!!

HAHAHAHAHAHAHAHA

I would NEVER use this in a serious discussion. HONEST!!

kuhnkat–
Yes. Deep climate in particular has been complaining about UAH. But it’s RSS that has the more negative trends this century. On the other hand, RSS has the higher long term trend.

Veterans Day Haiku by Andrew

Born of Liberty
United States of America persist
Thank you Veterans

Lucia: “next month, I’m planning to add the IPCC projections based on the models driven by A1B to the graph.”

I think you need a pretty low climate sensitivity for a good fit.

My fit of the average of UAH and RSS satellite temperature data using my 3 box model seems to match the recent plateau in temperatures fairly well despite using GISS forcing parameters:
http://moderateclimate.blogspot.com/

The corresponding climate sensitivity is pretty low: 1.2K

Steve–
You have a good case. But it seems some are convinced the data are eventually going to catch up with the models. All we can do is compare the data each month and wait to see, right?

“All we can do is compare the data each month and wait to see, right?”

Of course, and I would not be surprised to see a climate sensitivity of 2 K being shown the likely value, but the models with a sensitivity greater than 3 K are looking pretty questionable based on the data so far.

“wait to see”

It’s almost like nobody knows what’s going to happen.

Andrew

“A statistically significant cooling trend in RSS and UAH satellite data”.

I would have thought that belonged in the “Cherry Picking” thread.

bugs–
Well… I don’t use the same titles over and over.

bugs (Comment#23309) November 11th, 2009 at 10:42 pm

There are two reasons it’s not cherry picking Bugs. The first is that if you re-anomalize the post 2002 data the significance goes up quite a bit and occurs at more than one point. But the second reason is that the point of the original post was that RSS is having some substantial problems in trend.

I think there is a good case for re-anomalization is made by the fact that RSS has lost it’s balancing satellite (there is usually a day and night one) and UAH has switched to a station keeping one. In addition it’s backed up by a fourier transform of the data which shows strong annual signals in both post 2002.

I think the Cherry picking comment misses the point of the analysis.

Jeff–
It’s not a cherry pick to say UAH and RSS differ. That’s a pretty strong signal. And the difference between the two satellites was your major point.

I haven’t checked on re-anomalizing data after the instruments changed. You have a good point it should be done. I guess given the shortened baseline period, I’d probably also want to figure out whether we need to adjust the numbers of degrees of freedom to deal with the fact that re-anomalizing based on the period analyzed does always reduce “noise” a bit. (You know… sort of the reason why you need to subtract 1 from number of degrees of freedom when estimating the standard deviation by the sample standard deviation. There is always a slight problem with using the data to estimate an average and then subtracting that average.)

Also, in the context of the original post, I point out that it was discovered by accident while looking at the satellite transitioin, I showed both insignificant UAH and significant RSS over the same time period and followed it with the point that RSS is the measurement in error.

In another surprise post this morning I found that HadCRUT touched the significance line in the same timeframe. My post today was actually a tease for RC and Tammie inspired into action by your plot above.

http://noconsensus.wordpress.c.....een-years/

Steve Reynolds (Comment#23303),

Most of the models diagnose sensitivity of 3 +/- 1K per doubling. Since the current total radiative forcing (including CO2, methane, N2O, and tropospheric ozone) is 75% or so of the forcing from doubling CO2 from the pre-industrial level, the only way to justify 3K sensitivity is with a combination of substantial heat loss to the deep ocean combined with off-setting of ~1/3 of the radiative forcing by albedo increases from atmospheric aerosols. Since ‘global brightening’ since the early 1990’s from falling aerosol concentrations is pretty well documented, it is difficult to see how aerosol off-sets can continue to be used as the fig leaf that hides the models’… how shall we say.. ‘deficiencies’.

I have thought a bit about multi-box models as well. It seems to me that any physically meaningful model needs to have four boxes to allow absorption and transport of heat in physically realistic ways: atmosphere, land, upper ocean, and deeper ocean. You could treat land as a very fast box that sits side-by-side with the faster of the two ocean boxes, since solar energy falls on both (with the majority of solar energy absorbed by the faster of the two ocean boxes). The atmosphere would be a very low heat capacity (very-fast) box that sits on top of both the land and upper ocean box, but which does not itself absorb much solar energy. The atmosphere would serve to transport heat between the top ocean box and the land box, as well as out to space. Transport of energy between the faster ocean box and the deep (slow) ocean box would be as you suggest, controlled by a transfer constant.

SteveF,

The problem with a 1D four box model is that the structure and location of the land is important. Divide up the continents into islands surrounded by water and the temperature distribution and the average temperature will look quite different than if there were a single supercontinent. Central Asia at high latitude has a much bigger seasonal temperature range than Canada at high latitude. North, et.al. published on this in 1983: http://www.agu.org/pubs/crossr.....6576.shtml .

Jeff Id,

So which is it, you are headlining a ‘cooling trend’ or you aren’t doing that? Your response to the question of cherry-picking suggests that you were making a different point, but your latest post sure does seem to be doing the former.

Make your mind up, then if you decide you are headlining a cooling trend (as you are) we can expect you to respond to questions as to whether you’re cherry-picking or not. Of course, if you say you’re making some other point then you’d best go back and rewrite your latest post .

Simon–
I think the point of Jeff’s most recent point is that if you apply the same standard to decreeing statistically significant warming as we apply to statistically significant cooling, then you have to go back darn far in time to get ’statisically significant warming’.

This doesn’t mean that there has been no warming for 11 years. It doesn’t mean that the realization of weather we see is not associated with a process that would have an expected outcome of “warming”. All it means is that, when using a statistical razor, if that’s all the data we had, we would have to decree that the warming over 11 years is not statistically significant.

The thing is: The logic goes both ways. And from time to time “some” have tried to suggest that we “know” recent cooling cannot be associated with the hypothesis of long long term cooling because the short term cooling trend is not “statistically signficant”.

But that argument is false. Lack of statistical significance of data alone does not prove any hypothesis wrong.

SteveF,

is difficult to see how aerosol off-sets can continue to be used as the fig leaf that hides the models’… how shall we say.. ‘deficiencies’

Hey, it’s a big fig leaf, don’t underestimate it

Lucia,

I understand that. But my point is that you can’t duck out of a reasonable question about the impact of start dates etc. (c-p alert) by saying that you really weren’t talking about cooling trends, then next post the following opening paragraph -

The recent cooling shown in most global temperature datasets except GISS has been a spot of contention for the AGW believers. Global cooling has been an embarassing topic for those who have made careers of predicting worse than we thought warming every few weeks. This is especially contentious point for scientists who’s careers depend on warming.

That’s the rhetorical impact of words at work there – and full of all the crappy insinuation of motives stuff (“careers depend on”, etc.) .

I’m fine with a cooling trend, warming not significant, whatever. That would please me! I’m just questioning why Jeff says his ‘point’ was not cooling then makes clear in his next post that his ‘point’ is, er, cooling . Oh, and, er, the questioning of motives….

SteveF and DeWitt Payne,

I think you are both correct that adding more boxes could make the model somewhat more accurate, and adding a few more boxes to this type of time-step model is not very difficult. I may try some of your suggestions at some point.

However, the 3 box model may include land effects a little better than you think. Presently, the atmosphere box has a heat capacity about 2x that of dry air (this is what fits the data). I believe the higher heat capacity takes into account two additional effects:
1. the energy used to evaporate moisture (both from ocean and non-desert land including biomass) to keep relative humidity somewhat constant.
2. the heat capacity of the land biomass (trees, grass, crops…) and probably a few cm of depth into the land (several meters would be required to be major portion of the total atmosphere heat capacity).

The next couple of meters of land depth can probably be considered to be lumped into the upper ocean box (again likely a small effect relative to 100 m of ocean). Deeper depth into the land would be lumped into the deep ocean box, again with very small effect.

The point of all the above is that the 3 box model may not be too bad at including land effects if you are willing to consider biomass as part of the atmosphere box.

DeWitt Payne (Comment#23337),

There is no doubt that the land/ocean distribution is an important issue for a completely realistic model. For sure bigger land masses have bigger short term and seasonal temperture swings, since their area is (on average) further from the moderating influence of the oceans.

However, heat transport between the “land” and the upper “ocean” boxes would be described by a single constant in a four-box model; correct selection of that constant ought to give a better overall picture of the dynamic behavior than would a model where land is not treated as a separate box.

Simon Evans (Comment#23343),

“Hey, it’s a big fig leaf, don’t underestimate it”

But it is clearly known that aerosol concentrations are “shrinking”. The modelers need to be careful about how much they assume it hides.

Simon,

You claim to be an intellectual yet can’t seem to disaggregate between two different posts. The point of the second post was to show that the significance of trends cuts both ways, the first was errors in RSS which were strong enough to demonstrate ’significance’ in cooling.

And it’s quite irritating that you claim I ducked out of answering you. I most certainly didn’t “duck out” of your cherry picking question. Instead I pointed out how you completely entirely missed the point of the post and gave a reply which should have fully addressed your concerns. I wonder if you even bothered to read my post or did you just infer everything….. Sophistry?

You say — I’m fine with a cooling trend, warming not significant, whatever. That would please me! I’m just questioning why Jeff says his ‘point’ was not cooling then makes clear in his next post that his ‘point’ is, er, cooling . Oh, and, er, the questioning of motives….

You have got to have your head so far in the sand to not realize people have motives in global warming science. Many scientists have pointed it out including several famous ones. Careers do depend on the exaggeration of warming and people have even quit the IPCC over the bias. The examples are endless. Read Real Climate for a half hour and you can see it in full view.

I didn’t politicize global warming but I’m not slow witted enough to pretend it doesn’t exist and not write about it. Welcome to the Air Vent.

Jeff Id (Comment#23367)

Simon,

You claim to be an intellectual yet can’t seem to disaggregate between two different posts.

Where did I ever claim to be anything, Jeff? I have never appealed to my own authority (nothing to appeal to anyway!), though I’m happy to appeal to my own ignorance when that’s the case.

And it’s quite irritating that you claim I ducked out of answering you. I most certainly didn’t “duck out” of your cherry picking question.

It wasn’t my question, Jeff – look again.

You have got to have your head so far in the sand to not realize people have motives in global warming science.

Do you have motives, Jeff? Does Steve McIntyre? Roy Spencer? Richard Lindzen? Do you question any of their science on the basis of their presumed motives? Or is motive-questioning something you indulge in only with respect to those with whom you disagree?

I question the motives of someone who can’t seem to follow logic. There are two posts with two points both of which you seem to miss.

Lol! A revealing avoidance of the question, Jeff! Speaks for itself, I think.

Simon, I wasn’t attempting to answer your question. I was pointing out that you have insinuated the meaning of both posts incorrectly and then requested me to answer some off topic rubbish. Your half of the conversation has been nothing but sophistry (look up the meaning) and I don’t typically respond to sophistry.

Now we’re pooping all over Lucia’s nice blog so you can tell me I’m wrong again and have the last word.

I know you weren’t attempting to answer my question, Jeff, that is what “avoidance of the question” means (look it up if you have any trouble with the words – or better still, let’s avoid such puerile comments altogether, eh?).

I was pointing out that you have insinuated the meaning of both posts incorrectly and then requested me to answer some off topic rubbish.

Your second post, to which you linked here, includes your insinuations as to motives. How can my question be ‘off topic’ when you have linked to that post?

I can see that you won’t give a straight answer, so I won’t waste any further time giving you the opportunity to do so. I will maintain the working assumption that you are someone who likes to question the motives only of those with whom you disagree. Quelle surprise.