A few posts ago, I used the term “‘fingerprint'” (in quotations) in connection with the big red spot that one commonly sees in the upper tropical troposphere in contour plots of projected temperature trends by latitude. On an earlier occasion, we’d talked about Ross McKitrick’s “T3” ( here.)
A few readers contested my use of the term “fingerprint” in this context. “Fingerprints” are not a topic that I’ve covered at the blog so far nor can I say that I fully understand the operational usage of this term by climate scientists.
I get the impression that this is local dialect in climate science for something already known by a different name in conventional statistics. But without being able to work through examples from beginning to end, I find it hard to say what it corresponds to exactly. (And, of course, because it’s climate science, there’s nothing remotely resembling fully articulated examples to assist translation out of the local dialect.) Some day, I’ll try to figure it out.
IPCC TAR has a section entitled 12.4.3 Optimal Fingerprint Methods in which they refer to “optimal detection” (another term in the local dialect). For example, :
optimal detection studies of surface temperature have been extended…
Elsewhere IPCC TAR had a very peculiar appendix entitled: “Optimal Detection is Regression”, in which they say:
The detection technique that has been used in most “optimal detection” studies performed to date has several equivalent representations (Hegerl and North, 1997; Zwiers, 1999). It has recently been recognised that it can be cast as a multiple regression problem with respect to generalised least squares (Allen and Tett, 1999; see also Hasselmann, 1993, 1997)
Do “optimal fingerprints” also have something to do with multiple regression? It seems quite likely. It seems very odd that climate scientists are simply carrying out multiple regression with generalized least squares using strange and overblown terminology like “optimal detection” and “optimal fingerprints” without referring to the extensive pre-existing literature, but that seems to be what IPCC TAR says.
Lucia has a long and interesting thread on whether or not the T3 red spot is a “fingerprint” of AGW; I refer interested readers to her comments, with some of her readers contesting the idea that the red spot is an AGW fingerprint or that the seeming absence of the red spot “matters”.
For reference, here is an image shown last year at RC showing the effect of 2xCO2. (The image is entitled “2xCO2_tropical_enhance.gif”.
.
Figure 1. Effect of 2xCO2 (from RC).
Gavin also showed a similar T3 hot spot with large solar forcing, observing:
If the pictures are very similar despite the different forcings that implies that the pattern really has nothing to do with greenhouse gas changes, but is a more fundamental response to warming (however caused). Indeed, there is a clear physical reason why this is the case – the increase in water vapour as surface air temperature rises causes a change in the moist-adiabatic lapse rate (the decrease of temperature with height) such that the surface to mid-tropospheric gradient decreases with increasing temperature (i.e. it warms faster aloft). This is something seen in many observations and over many timescales, and is not something unique to climate models.
As a little exercise, I downloaded monthly data from UAH, made zonal averages in 10-degree bands and calculated trends for three levels (T2LT, T2 and T4) and likewise for CRU, and then made a contour plot, using altitudes of 2.5 km (740 hPA), 6.1 km (466 hPa) and 20 km (75 hPa). This resulted in the following graphic:
Figure 2. Contour of trends using CRU and UAH. (deg C/year)
Re-examining the matter, I noticed two quotes in AR4 that seem relevant (and which haven’t been mentioned at Lucia’s yet). AR4 Box 8.1 states:
GCMs find enhanced warming in the tropical upper troposphere, due to changes in the lapse rate (see Section 9.4.4).
and again:
At low latitudes, GCMs show negative lapse rate feedback because of their tendency towards a moist adiabatic lapse rate, producing amplified warming aloft.
If these are relevant quotes, then the presence/absence of Clifford the Big Red Dog might well be related more to whether water vapor/cloud impact has been adequately modeled, as opposed to the direct GHG effect.. If this interpretation is valid, then the point made by Gavin and others would seem to be an empty rhetorical point, as it seems to me that there has been more concern over the “multiplier” effect of water vapor/clouds than over the direct GHG effect. So the point remains worth discussing whether or not it is a GHG “fingerprint”. The water vapor/cloud feedbacks are very bit as relevant to understanding the matter.
As discussed on many occasions, it’s quite possible that the Big Red Dog is really there and the issue is with UAH algorithms (at some point, I’ll try to decode RSS gridcell data and do the same plot from their data.)
Having said that, the pattern, as shown above, has both points in common and that differ from the models. Stratosphere cooling and an Arctic hot spot are in common; but, in addition to the seeming absence of the Big Red Dog, there seems to be much more noticeable NH-SH asymmetry in the observations than in the models.
UPDATE: Here’s the corresponding graphic for RSS (which uses 5 levels – apparently there are some issues with the lower stratosphere values but I can’t comment on this. )
Figure 2. RSS Trends. (deg C/year)
There seems to be a noticeable gradient in stratosphere trends as shown below. I don’t see anything matching this in Gavin’s Big Red. Has anyone seen this discussed?
Climate Audit 
191 Comments
#seems to be S to N, dateline to Dateline
#TLS is centered around 75 hPa (~20 km or ~14-50 km) and
#TTS (channel 3) around 220 hPa (~10 km or 1-35 km).
#Douglass: T2LT mean altitude 2.5 km (about 740 hPA);
#T2 – mean altitude 6.1 km) 466 hPa
#http://www.sensorsone.co.uk/altitude-pressure-units-conversion.html
#make sure you install the R packages ‘ncdf’ and ‘fields’ in the R-workspace before running this script or these next two lines will fail and you’ll feel like an idiot….
library(ncdf)
library(fields)
hpa=c(1000,740,466,75)
lat=seq(-85,85,10)
alt=c(0,2.5,6.1,20)
####MAKE 3 TIME SERIES OF ZONAL SAT INFO (10 degree bands) AND ONE SURFACE
prefix=c(“http://vortex.nsstc.uah.edu/data/msu/t2lt/tltmonamg.”,
“http://vortex.nsstc.uah.edu/data/msu/t2/tmtmonamg.”,
“http://vortex.nsstc.uah.edu/data/msu/t4/tlsmonamg.”)
suffix=c(“_5.2″,”_5.1″,”_5.1”)
zonal=rep(list(NA),4)
names(prefix)=names(suffix)=c(“t2lt”,”t2″,”t4″)
names(zonal)=c(“surf”,”t2lt”,”t2″,”t4″)
year=1979:2008
index=seq(1,1+ 649*11,649);index
#this keeps trak of index lines in MSU record
#input 3 satellite levels
for(k in 1:3) {
sat=ts(array(NA,dim=c(length(year)*12,10368)),start=c(1979,1),freq=12)
for(i in year) {
loc=paste(prefix[k],i,suffix[k],sep=””)
fred=readLines(loc)
notindex=is.na(match(1:length(fred),index))
writeLines(fred[notindex],”temp.dat”)
if(k==3) {x=read.fwf(“temp.dat”,widths=rep(5,16)); x=c(t(x))} else {
x=scan(“temp.dat”,n=10368*12)} #124416
n=length(x)/10368
sat[(i-1979)*12 +(1:n),]= t(array(x,dim=c(10368,n)) )
}
N=nrow(sat)
zonal[[k+1]]=ts(array(NA,dim=c(N,18) ),start= c(1979,1),freq=12)
for (j in 1:18) {
temp= 576*(j-1)+ (1:576)
zonal[[k+1]][,j]=apply(sat[,temp],1,mean,na.rm=T)/100
}
}
#input surface
##library(ncdf)
download.file(“http://hadobs.metoffice.com/hadcrut3/data/HadCRUT3.nc”,”temp.dat”,mode=”wb”)
v=open.ncdf(“temp.dat”)
instr=get.var.ncdf( v, v$var[[1]]) # 1850 2006
dim(instr)# [1] 72 36 1899
#this is organized in 72 longitudes from -177.5 to 177.5 and 36 latitudes from -87.5 to 87.5
dim0=dim(instr)
instr=aperm(instr,c(3,1,2)) #
instr=array(instr,dim=c(dim(instr)[1],dim(instr)[2]*dim(instr)[3]))
zonal[[1]]=ts( array(NA,dim=c(dim(instr)[1],18)),start=c(1850,1),freq=12)
for (j in 1:18) zonal[[1]][,j]= apply( instr[,144*(j-1)+(1:144)],1,mean,na.rm=T)
#rm(v);rm(instr)
# zonal[[1]]=zonal.rss[[1]]
save(zonal,file=”d:/climate/data/satellite/zonal.msu.tab”)
###CALCULATE TRENDS
Trend= array(NA,dim=c( 4,18))
for (k in 1:4) {
year=c(time(zonal[[k]]))
temp=(year>=1979)
for (j in 1:18) {
fm=lm(zonal[[k]][temp,j]~year[temp])
Trend[k,j]=fm$coef[2]
}
}
##PLOT CONTOUR MAP
#library(fields)
hpa=c(1000,740,466,75)
lat=seq(-85,85,10)
breaks0=seq(-.055,.055,.01);n=length(breaks0)
filled.contour(x=lat,y=1000-hpa,z=t(Trend),levels=breaks0,col=tim.colors(n-1),
main=”UAH Trends”, plot.axes = { axis(1, seq(-60, 60, by = 20))
axis(2, at=1000-hpa,labels=as.character(hpa)) } )
You can see Mann’s fingerprint here.
I’m curious if any of these IPCC guys understand that the word “optimal” has very specific meaning and absolutely must be referred to some specific criteria, e.g., minimum mean square error, else is it nothing more than hype.
Mark
I’ve seen the IPCC modelled warming trend plot for some time. But I’ve never heard a reasonable explanation as to why the surface temperature trend at the South Pole is “grayed out”. Anyone have insight into this?
Re: schlew (#4)
Central Antarctica has high elevation (3000+feet). The greyed out region is underground.
Re: Raven (#9),
Raven – Thanks, I didn’t realize that
Steve,
If I read your scale correctly, it only ranges between -0.05 C and +0.05 C. Using the same color scale as Gavin’s would make your figure entirely white. Correct?
Is the 2xCO2 graph linear in relation to CO2? In other words would 1.5xCO2 just be half of the values presented? I know that an argument has been made for a log relationship but that doesn’t include other forcings. What is the expected value based on the time frame you used?
Steve
I get errors running your script:
and later
Since we’re at the top of a thread, I think it’s worth mentioning that Pat Keating (aka PatK ) published Simple radiative models for surface warming and upper-troposphere cooling, a simple model that extends the widely known analysis predicting the adiabatic lapse rate. By partitioning photons and sections of the atmosphere, he develops a theoretical models that predicts now hot spot. Near the end of the paper, he discusses how some key analyitical (over)simplifcations may be causing the hot spot to be predicted though the opposite effect may occur on earth.
FWIW– The IPCC sometimes puts “fingerprints” in scarequotes too! As far as I can tell, this word is used all sorts of way. Why people want to jump into threads and say “You can’t use it to mean ‘X'” when numerous examples of that usage exist in the peer reviewed literature, newspaper interviews etc. The IPCC glossary definition is pretty broad.
Now… since my blog seems to have clocked at least 300 comments on this, I’m going to go watch football and crochet an afghan. ( UK commenters…. I think you would call this a shawl or a rug. )
Actually the first part works so long as you set up your CRAN repository correctly *headslap*
The second part doesn’t work. I think you’ve missed a definition somewhere.
#7. I re-executed the script and have inserted a slightly tidied version which shouldn’t hang up.
Steve Mc said:
“If these are relevant quotes, then the presence/absence of Clifford the Big Red Dog might well be related more to whether water vapor/cloud impact has been adequately modeled, as opposed to the direct GHG effect.”
Their “hot spot” is for all GHG’s and water vapor is considered a GHG by them. That IS how they manage to “find” 3c+/century. I suggest that they have NOT modeled the feedbacks correctly per Dr. Spencer’s work. I would also suggest their their dubious reanalysis of the ‘sonde data trying to “prove” the hot spot that isn’t there that invalidates their model projections is simply desperation to “SAVE” their mistakes.
Now we have obfuscation in the camp trying to redefine the MODEL results, without which, someone(s) in the warmer camp needs to present a lot of new science to support the BELIEF in impending disaster.
Please note that the “hot spot”, which is a faster warming of the tropical troposphere than the tropical surface, must be accompanied by an elevation of the tropopause and cooling of the stratosphere to be what the models show. The ratios between these are probably also important as they alledgedly represent particular interactions between convection, conduction, radiation, mass, and other physical characteristics in our atmosphere.
Lucia,
you are going to crochet a dog or a native of Afghanistan??
There seems to be a noticeable gradient in stratosphere trends as shown below. I don’t see anything matching this in Gavin’s Big Red. Has anyone seen this discussed?
#12. Afghanistan. I traveled overland over the Khyber Pass when I was young – not a journey that I’d care to do today.
I’ve read it over a few times, and still can’t see how Gavin’s comment (that a T3 red spot is not a fingerprint of AGW in particular but rather of warming in general whatever the cause) can be taken to be “empty rhetoric”. Is it just because he’s wrong? If so (and if I have misunderstood the rationale, please let me know) isn’t that ok? I mean, at least he uttered a statement that COULD be wrong, something perhaps that has not always been the case in the past.
#12. IT seems to me that elevation of the tropopause is something that would tend to mitigate the impact of 2xCO2 on the surface. So while it would be evidence of CO2 impact, it wouldn’t seem all that hard to adapt to a tropopause that was 150 m higher.
#15. maybe I didn’t express this as clearly as I might have. What I had in mind was this: Gavin seemed particularly concerned about saying that Big Red was not necessarily linked to CO2 increases. If that point were stipulated for the sake of argument, then Big Red absence would presumably indicate a problem with the water vapor feedbacks which are just as much at issue. So Gavin’s rhetorical point wouldn’t get him anywhere in terms of vindicating Big Red. I’m not trying to over-state things here.
Here is RSS script:
#TLS is centered around 75 hPa (~20 km or ~14-50 km) and
#TTS (channel 3) around 220 hPa (~10 km or 1-35 km).
#Douglass: T2LT mean altitude 2.5 km (about 740 hPA);
#T2 – mean altitude 6.1 km) 466 hPa
library(ncdf)
library(fields)
hpa=c(1000,740,466,220,75)
lat=seq(-85,85,10)
alt=c(0,2.5,6.1,10,20)
zonal=rep(list(NA),5)
loc=c(“http://www.remss.com/data/msu/data/netcdf/RSS_Tb_Anom_Maps_ch_TLT_V3_2.nc”,
“http://www.remss.com/data/msu/data/netcdf/RSS_Tb_Anom_Maps_ch_TMT_V3_2.nc”,
“http://www.remss.com/data/msu/data/netcdf/RSS_Tb_Anom_Maps_ch_TTS_V3_2.nc”,
“http://www.remss.com/data/msu/data/netcdf/RSS_Tb_Maps_ch_TLS_V3_2.nc”)
for(k in 1:4) {
download.file(loc[k],”temp.dat”,mode=”wb”);
v< -open.ncdf("temp.dat")
#V$longitude 2.5 deg ells Greenwich to Greenwich
#v$latitude 2.5 deg from S to N
# v$vals 0 to 371 from Jan 1978
instr <- get.var.ncdf( v, v$var[[1]]) #
dim(instr)# [1] 144 72 372
range(instr) # -9999.00000 11.66557
temp=(instr< -9000)
instr[temp]=NA
lat=v$dim$latitude$vals;
instr=aperm(instr,c(3,1,2)) # 372 144 72
A=array(instr,dim=c(dim(instr)[1],dim(instr)[2]*dim(instr)[3]))
A=A[13:nrow(A),] #start in 1979,1
zonal[[k+1]]=ts( array(NA,dim=c(dim(A)[1],18)),start=c(1979,1),freq=12)
for (j in 1:18) zonal[[k+1]][,j]= apply( A[,576*(j-1)+(1:576)],1,mean,na.rm=T)
rm(v);rm(instr);rm(A)
}
#input surface
##library(ncdf)
download.file("http://hadobs.metoffice.com/hadcrut3/data/HadCRUT3.nc","temp.dat",mode="wb")
v<-open.ncdf("temp.dat")
instr =1979)
for (j in 1:18) {
fm= try(lm(zonal[[k]][temp,j]~year[temp]))
if (!(class(fm)==”try-error”)) Trend[k,j]=fm$coef[2]
}
}
##PLOT CONTOUR MAP
#library(fields)
hpa=c(1000,740,466,220,75)
range(Trend,na.rm=T) #-0.08826997 0.05186060
breaks0=c(-0.095,seq(-.055,.055,.01),.095);n=length(breaks0)
filled.contour(x=lat,y=1000-hpa,z=t(Trend),levels=breaks0,col=tim.colors(n-1),main=”RSS/CRU Trends”,
plot.axes = { axis(1, seq(-60, 60, by = 20))
axis(2, at=1000-hpa,labels=as.character(hpa)) } )
KuhnKat–
In answer to your question, the afghan is neither a dog nor an Afghani.
As you can see, Having one of these draped over a couch is a distinctive pattern sometimes called the “fingerprint” of “Americana”. Having five or more draped over the couch is the fingerprint of trying to keep the heating bills low. (Calling it an afghan is also a fingerprint of being American. The word seems to confuse people from the UK. I have no idea what Canadians call these. )
Looking at the GISS image, I’m tempted to design one to show the TT hotspot.
Re: lucia (#19),
Looks like the fingerprint of a Goodyear radial.
Re: lucia (#19),
OT
This side of the pond ‘An Afghan’, to my generation, meant an Afghan Coat
http://www.peterloud.co.uk/photos/Afghanistan/Af_6.html
or maybe a hound. (Quite big but not usually Red to my recollection.)
So I ran a quick search and found this site with some superb photos, notably from Afghanistan and India in the early 70s – a time Steve is probably familiar with …. a post seemed called for even though OT.
Gavin’s post in December 2007 was to counter things like this:
http://www.climateaudit.org/?p=1700
And this:
And this:
http://www.theaustralian.news.com.au/story/0,25197,24036736-7583,00.html
And this:
http://scienceandpublicpolicy.org/monckton_papers/greenhouse_warming_what_greenhouse_warming_.html
i.e. no hotspot = no enhanced greenhouse effect = climate change natural. I also don’t see any quotation marks around words like “fingerprint” and “signature”.
And whenever satellite “channels” are discussed, it must always be mentioned that they are weighted averages of large swaths of atmosphere. The temperatures represented by one “channel” don’t end where the next “channel” begins, which is why the documentation says things like
Click to access Mears_and_Wentz_TMT_TTS_TLS_submitted.pdf
Also, I don’t know how much it affects your plot, but RSS TTS begins in 1987, not 1979.
cce, or was the purpose of Gavin’s post a desparate distraction measure? Face it, we can see temperature trends in the atmosphere and they do not add up to what the models predict.
Re: Mike C (#23),
I “see” trends that encompass huge swaths of atmosphere stirred up in a blender.
To compare the satellite “channels” with equivalently weighted radiosonde analyses, go to http://www.remss.com/msu/msu_data_validation.html#compare, choose “summary” and then choose the channel in question.
As far as what the models predict and what the data tell us, these papers may be of interest:
Click to access sondeanal.pdf
http://www.nature.com/ngeo/journal/v1/n6/abs/ngeo208.html
Click to access VinnikovEtAlTempTrends2005JD006392.pdf
Click to access i1520-0442-21-18-4587.pdf
Re: 24-27. None of the quotes provided are inconsistent with what has been said. The GCMs do expect a tropical “hotspot” to emerge, regardless of the cause of the warming. As far as a “material difference”, you may calculate if the discrepencies between the two time periods creates a material difference or not.
Ok, let’s try again. How do statements by Gavin Schmidt et al 2005 square with what is being discussed now? Or has it since been abandoned?
Click to access 881407.PDF
#22. As quoted above, AR4 says:
thanks for pointing that out. But what relevance does this have to the trends for the purpose of this plot? Do you believe that the 1979-86 interval makes a material difference?
#24. Santer (Schmidt) et al begins:
That sure sounds like Clifford the Big Red Dog. Maybe they’ve “moved on”.
Some AGW defenders claim that the absence of the tropical tropospheric hotspot (aka the Big Red Spot) in satellite data is because the tropospheric temperatures contain some of the stratospheric cooling. While it is true that the raw satellite “channel” data do contain temperatures over a large range of altitudes, the providers of the satellite data also claim to be able extract actual altitude temperatures.
My point is that if they were failing to remove stratospheric cooling from the troposphere, such that the tropical Big Red Spot was being masked by cooling, then we would ALSO see tropospheric cooling OUTSIDE the tropics. But we don’t. Indeed, the UAH data even seems to show evidence of slightly greater tropospheric cooling in the tropics than out of it…
Re: Chris H (#29),
They make no such claim. The closest they come is the lower troposphere analyses.
I will also point again to RSS’ page which allows an apples to apples comparison between radiosonde analyses and satellite “channels.” The analysis which shows the most tropical troposphere warming, RAOBCARB 1.4 (green Xs), shows even lower warming than RSS (triangles) when weighted the same.
http://www.remss.com/msu/msu_data_validation.html#compare
Of particular interest:
BTW, when I say “slightly greater tropospheric cooling in the tropics than out of it”, what I really meant was that the the troposphere looks slightly cooler in the tropics than out of it.
Steve
Got a lot further as the script reads a lot of files from the website. Then it blows up:
You’ve picked up an extra space between v and open courtesy of WordPress. Remove it and it should work. Or use
IF (and this is unclear to me) the big red spot can be caused by any sort of warming, and it is NOT found, then either 1) it has not warmed enough over the satellite era for the big red to be visible or 2) much of the surface warming is due to land use change (per Pielke sr) and bias in instruments (UHI) per Watts. Big red is a quantitative phenom, not qualitative, and there is noise in the data, so it is hard to say it was not found, which of course gets us back to the whole Dougass vs Santer debates.
cce–
Gavin’s Dec post was to counter Ross? Or SteveM’s post? Are you sure? Why doesn’t Gavin mention Ross or Steve? Ross’ tax proposal? The structure of Gavin’s article would appear to be a rebuttal of Douglas. PERIOD.
If you are correct, and Gavin was actually rebutting article he doesn’t mention, and advancing an argument his readers might not have read then what we learn is: Gavin has latched onto a technique guaranteed to make his argument obscure, incoherent and unpersuasive.
Re: lucia (#34),
Douglass et. al. does not claim that the hotspot is unique to greenhouse warming, unlike the statements that I quoted. So when Gavin says,
. . . he is addressing the oft-repeated belief that only an enhanced greenhouse effect creates the “hotspot.”
If you regress the monthly TMT anomalies against the TLT anomalies for the same month for tropics land only (I haven’t tried land and sea or sea only) for either UAH or RSS, the slope is less than one with a confidence interval that doesn’t include one. I’ve also done multiple regression on TMT against TLT and TLS. The fit is worse and the slope of TMT against TLS is ~0.025 so the contribution of TLS to TMT is quite small. A slope of less than one implies to me that TMT is warming at a slower rather than a faster rate compared to TLT. I think this is a better test than comparing OLS linear time series trends because the linear trend has a large error that is made larger by the reduction in dof caused by autocorrelation.
Is this a valid technique? Does it mean what I think it does? Am I missing something like autocorrelation that would make the test less significant?
#15: Chopbox, some of the earlier discussion on the tropical troposphere thread fills in essential background, including the reference to IPCC Figure 10.7 and IPCC Figure 9.1. The combination of those diagrams implies that in response to observed historical changes in the usual forcings (a key qualifier) only GHG’s would produce a tropical tropospheric redspot, and it should be observable now. The rejoinder from some modelers is that increased solar output would also, in principle, produce a similar hotspot, so it’s not a “unique” fingerprint of GHG’s. Realclimate even produced a diagram to show it, which looks impressive except that to get the effect they fed in a 2% increase in solar flux, even though the IPCC claimed total irradiance had risen only 0.04% since the end of the Maunder Minimum.
By way of analogy: I have a bucket of red paint (GHG’s) and a bucket of yellow paint (sun) and the wall is blue. Which bucket didn’t I use? Answer: we can unambiguously say both. We can refine the answer if we also know that the wall was supposed to be red. Then we specifically know I didn’t use the red paint. But that’s not necessary. The rejoinder says: if the wall was red and both buckets contained red paint then you wouldn’t be able to tell.
To spot the flaw in the attempts to downplay the hotspot issue you have to watch when people flit between hypothetical and actual worlds. The potential non-uniqueness would matter if the hotspot were there and solar output had gone up a lot and the climate were more sensitive to solar change than the IPCC has otherwise claimed. That’s a hypothetical world. But in the observed world, GHG’s went up, the hotspot is not there, and the sun has (apparently) not intensified much. That’s a problem for the GHG story.
Regarding the logic of the T3 tax, we expect the sun primarily to cycle rather than trend strongly up or down. But we expect GHG’s to primarily trend up rather than cycle. So if the tax is calibrated to the tropical hotspot, even if both solar and GHG’s “matter” a trend in the tax will primarily reflect the influence of the GHG’s. Solar-induced increases will tend to average back out over the decades. Going into the details a bit, it’s OK if the tax rate rises due to increased solar effects if GHG’s cause higher marginal damages during intervals of higher solar output, which in principle makes sense. More details are here.
Re: Ross McKitrick (#36), Dear Sir, thank you for all your contributions. I find your economic ideas concerning attaching a carbon tax to the troposphere warming attractive if we are forced into a carbon tax. However the entire research appears to be focused on possible damages from increased CO2. Should the known benefits of increased CO2 be considered against the potential harm ? Have the financial consequences of these known benefits been quantified? Sorry to get off topic but you are the only one I know of who may be qualified to put a financial number on the saving of CO2 in regard to water use for food, increased vegetation becoming a carbon sink, energy savings on fertilizer, land use etc. Does the effect of aerial fertilization of the planet increase in a linear trend, while the warming effect of 2 x CO2 decreases logarithmically?
Steve: please do not introduce policy issues. They are important but excluded for editorial reasons.
Paul Linsay’s question back in #5 appears to have fallen by the wayside:
I’d guess that Gavin’s graph shows the total projected effect of 2XCO2, whereas Steve’s units are the rate of change per year in observed temperatures over the past 20 years or so (per #26) of satellite data. If so, then .05 on Steve’s scale would amount to about 1.0 on Gavin’s scale. Even so, Steve’s graph would lie almost entirely in Gavin’s white zone, as you note.
I gather Gavin has chosen his outer color bands at +3°C to +14.6°C and -3°C to -13.2°C in order to dramatize the extremity of his projection, which I suppose is fair enough. However, I find it odd that he has made the subdivisions quite unequal: The central white band extends 1.0 °C each size of 0, but then the next bands are respectively 0.8, 0.4, 0.3, and 0.5 wide, before the final bands in excess of 10. Is this standard practice in such graphs?
The unlabeled vertical axes appear from Gavin’s post to be in millibars, so that the vertical axis is indirectly measuring altitude, in informative units of fractions of the total mass of atmosphere.
There is an intriguing “Little Red Puppy” at the N pole in Gavin’s graph. However, his horizontal axis is unfortunately in degrees of latitude rather than cosine of latitude, so that his graph exaggerates the importance of the polar regions for the total volume of atmosphere.
Where can I get the “fields” and “ncdf” add on packages for R?
Where can I get the “fields” and “ncdf” add on packages for the Windows version of R?
Re: 2dogs (#38),
Boot up R. Click on Packages, then on Install Package(s). Choose a mirror (download site) in the box that appears followed by clicking on the package you wish to download. Repeat for each package.
#15: As I understand Gavin’s argument, he claims that the “red dog” results from water vapor feedbacks to rising temperatures and that any rise in temperature (no matter what the cause) should result in the same pattern.
Therefore, the absence of a red dog does not imply that something other than anthropogenic GHGs are heating the atmosphere.
This is empty rhetoric because the red dog is clearly predicted by climate models, so the absence of a red dog strongly argues against the accuracy of those models. In particular, calculated values of climate sensitivity to CO2 (James Hansen reports that climate scientists have nailed this value at 3 degrees centigrade) DEPEND on a fairly massive feedback from water vapor.
If anthropogenic green house gases are heating the atmosphere as predicted, but water vapor feedback is not happening as predicted, then the climate models are still invalid and the climate sensitivity calculations are still wrong.
—
I find myself agreeing with the current consensus right up to the point where they calculate the water vapor feedback, at which point I think they are drawing conclusions that are far too strong based on far too little data. In my view the absence of a red dog is a glaring problem that demands explanation.
In short, somebody needs to explain why we should trust climate models when the atmosphere behaves in a radically different way from what the climate models predict. Santer et al basically said said: “based on the subset of data we analyzed, it is still statistically plausible that we have had a run of bad luck, and the models are correct after all”. Even if that were correct, it is hardly a convincing argument.
Insolation varies appreciably during a year, as in summer and winter. It seems reasonable that the tropical upper troposphere should show hints of the Big Red Dog (at least some of the hair of the dog) during local summer. Does it?
For example, does the atmosphere above, say, Palau in the Western Pacific, full of rising parcels of moist air, show amplified warming in boreal summer, when compared to Palau in boreal winter?
And, for that matter, does the tropical upper troposphere in the descending regions of the Hadley-Walker circulation show changes in amplification from summer to winter? These regions of descending air make up the majority of the tropical atmosphere and therefore they’re important in any discussion of changes in the tropical atmosphere’s profile.
I’ve spot-checked soundings in Palau and elsewhere and, so far, I’ve seen little to no evidence of solar-driven seasonal amplification. Perhaps there’s a study somewhere which examines more data and the processes involved and can offer guidance on this.
There is also a question in my mind of how additional CO2 affects the radiational cooling of the very dry descending air parcels and thus the vertical temperature profile in the descending regions. That is something which would not be a factor in solar-driven warming.
#39 excellent summary from Jason. Even if Gavin’s explanation makes any physical sense, which I doubt, it is simply inconsistent with the observations. However you look at it, it’s wrong. The alleged increase in water vapour, so loved by Gavin and his fellow global warming exaggerators, is not observed, see
Smith, Yin and Gruber, Geophys. Res. Lett. 33, L06705 (2006),
Nedoluha et al, J. Geophys. Res. 108, NO. D13, 4391 (2003).
Both these papers show no significant trend in water vapour.
It is simply absurd to claim that these two warming mechanisms, one short-wave mostly absorbed at the surface and the other long-wave absorbed in the atmosphere, will lead to the same warming picture. I cannot understand why Gavin and his apologists like cce attempt to push this nonsense.
Re: PaulM (#41),
It’s wrong if you ignore observations that are inconsistent with your predetermined beliefs. See above for four examples. See below for another.
vs.
Click to access Soden05.pdf
The shortwave becomes longwave when it is emitted by the surface. There will be more outgoing longwave with an increase in solar.
With an enhanced greenhouse effect, the energy is slowed on its way to space. In both cases there is more energy in the troposphere and the result is the same.
Re: cce (#55),
This is hardly true. The emitted outgoing radiation depends on the temperature of the matter. Whether the surface is warmed by the Sun direct radiation or downward longwave GHG radiation, that has no consequence in the IR outgoing radiation.
I don’t know if it physically sounded to think of a possible differential effect on trends between day and night outgoing radiation in the IR window, but since troposphere is a very active region, not in radiation equilibrium, and the Earth surface is not a matter that simply transforms Sun radiation in sensible heat, you have to be very lucky to find out a result.
Jason # 39, I think Gavin was showing the big red dog plus a blue umbrella for GHG’s and just a big red dog for solar (keeop in mind he really had to turn up the solar to get the dog)
cce, naw, I think I’ve seen enough attempts at excuses on Lucia’s thread to tell that the AGW crowd is scared of this one. I’m not so sure about greenhouse warming in all this, but the enhancement is pretty much out of the picture.
When I look at the two graphs that Steve Mc put up, it starts to make sense… look at the vertical nature of the changes in temps… heat rises, need I say more?
Steve Mc, It would be real helpful if you could do the same with the baloons and and put the dates on your graphs.
The (absence of) big red dog ate my homework.
In Svensmark’s theory, the two poles act in opposite ways to cosmic rays. This has to do with the effect of clouds over an ocean vs clouds over a white continent acting in opposite ways. This is his polar see-saw. The graph in this post sure looks like that theory predicts.
Somebody should ask Gavin the philosophical question: “how many Red Dogs can dance on the head of a pin?”
Gavin says that 2000 was the coldest year in this decade. Anyone else see the problem with statement.
Re: Vernon (#47), And which decade would that be?
Well, I am guessing he intended to mean the current decade, but it goes from 2001 to 2010. But I am not a climate modeler.
Here’s a time series of tropical Pacific air temperature at various levels of the atmosphere. The region covered is 5N to 25N, from 110E (near Vietnam) to 100W (near Central America). The period is Jan 2003 to November 2008. The y-axis scaling is the same for all four series:
There are distinct solar-driven changes (seasons) apparent in the lower tropical troposphere. Their magnitude (peak to trough) is about 3C at 1000 mb. This seasonal pattern is less distinct but still visible in the mid-troposphere (500mb). At the Red Dog level (300mb) the seasonality is much less distinct and the magnitude of the changes are smaller, not larger, than those near the surface.
Admittedly this is a simplification of a complex matter and uses arguable (reanalysis) data. Also, a change in solar output would affect the global atmosphere over a longer time scale, which is not the same as a regional, seasonal change. Nevertheless, I’d expect to see some of the hair or maybe a paw of the Red Dog ( a hint of seasonal amplification in this data), if Sol can create the Dog.
Re: David Smith (#51),
Here’s a thought: If you do an annual average then it’s possible to equate pressure and altitude. On a shorter time scale, though, doesn’t the atmosphere expand and contract as it warms and cools? That would mean that the 850 mb altitude would go up as the temperature increased and down as it decreased. Since the lapse rate is a function that relates temperature to altitude not pressure, then an increase in altitude would mean a lower temperature at the same pressure, so the temperature curve at constant pressure is damped compared to the surface as altitude increases. Maybe. I think.
Does your data source have an humidity time series as well?
Steve
I now get past the previous syntax errors but I still get more substantial SNAFUs:
Inter alia I would like to point out that the line which saves to d:/climate/data/satellite/zonal.msu.tab won’t work if you don’t have a D drive or this directory structure.
More evidence of Dark Enthalpy. The hot spot is most likely there, substantial, and intact, it is just not detectable by instruments in this energy phase space.
Re #52 Thanks, DeWitt, interesting points. I’ll plot 200mb to 400 mb thickness (difference between heights) to check for seasonality, as thickness should be proportional to the temperature of that layer.
In any case the y-axis on the Red Dog (Figure 1) is millibars of pressure. I assume the y-axis remains constant across the figure. An extension of the 300 mb line across the figure shows a temperature increase of over 3C at 300mb at the equator versus 1.8C at the surface.
Regarding humidity, the source (reanalysis data) has relative and specific humidity time series, but those values are considered suspect. I’ll be glad to plot them, with that quality caveat, if you have something in mind.
Re #60 I get a 400mb -to-200mb thickness around 4800 meters. The thickness time series resembles the wiggles of the 300mb plot on the graph in #51. I’ll post the plot later.
Annual thickness peak-to-trough runs about 30 meters (0.6%). If a back-of-envelope PVT approximation is appropriate, then the 0.6% volume range indicates a peak-to-trough temperature range of about 1.5K in the layer, which looks reasonably close to the range shown by the 300mb temperature.
Re: David Smith (#61),
So the within layer altitude change would be about 15 meters. However, you have to include the expansion of the atmosphere below the 400 mb level as well. That’s three times as much and a bigger temperature change as well. That should be well over 100 meters or on the order of a 0.5 to 1.0 K reduction in the annual temperature range. I think the MSU’s measure as a function of pressure rather than altitude too. It seems too obvious, though. That sort of thing has to be built into the models as it’s Physical Meteorology 101.
Re: DeWitt Payne (#62),
To follow up: Using an exponential atmosphere approximation and a constant lapse rate of 6 K/km, the rate of temperature change at 400 mb pressure is indeed less than the rate of change at the surface. Here’s some numbers.
surface T altitude temperature@400mb
298.2 7.487 253.28
299.7 7.523 254.55
301.2 7.564 255.82
For a delta T surface (or at any constant altitude) of 1.5 K, the delta T at 400 mb is 1.27 K. Altitude is in km. A delta T of 3 K produces an altitude change of 77 m for the 400 mb pressure level.
There was a long debate over the hotspot at Club Troppo.
The Big Red Dog must be in the model results because of the supposed Big Water Vapour Feedback. When the Big Red Dog is absent from the actual temperature data, it means that the Big Water Vapour Feedback does not exist in reality. The models overestimate the water vapour feedback. In reality it is small or may not even exist at all. This conclusion does not depend on the cause of the warming.
cce: Thanks for the link to the Soden paper. How does the strength of the WV feedback relate to the overall sensitivity outcome in GCM runs? I.e., does the difference between 1.5K versus 4.5K sensitivity result from different assumptions about WV feedback? I would have liked some indication in the Soden article about where a constant relative humidity situation leaves us regarding the implied strength of the overall feedback.
I found their discussion of the spatial fit less convincing than the T2-T12 comparison. Figure 3 does not look like a good fit. This is a difference between economists and climatologists maybe. In my world you couldn’t print those jumbled up lines and make a qualitative comment that they match up if you squint just so, especially when the graph is a linear projection from a spherical world, so any mismatch over the tropics is downplayed because the region is compressed. You’d need to quantify it. There are areas where the GCM backcast does not follow the MSU, and I suspect a goodness-of-fit test would be dicey.
Re: Ross McKitrick (#65),
According to some new work by Roy Spencer, the water vapor and cloud feedback is all negative.
Re: jae (#70),
Click to access Dessler2008b.pdf
Re: cce (#87),
Do you actually read these papers that you refer to determine their credibility?
This paper looks at some data from the period 2003 to 2008, uses some hypothetical equations to calculate some parameters, and blandly presents their estimated values as factual. They look at their results and they “see” and “show” all sorts of conclusions and implications. There is not a single measure of uncertainty for any estimated parameter given in the paper! The closest they to get to statistical analysis is calculating an average. Some examples from the paper:
Given the small amount of data you might think that they might realize that these differences could be related to uncertainties in their data and not their inferred rationalizations. Hardly proof positive of anything.
Re: cce (#87),
Well, cce, Roy Spencer has specifically challenged those folks to show him where he’s wrong, providing his email address for that purpose. I sincerely hope that Dessler and the others that believe in positive water vapor feedback will take Spencer up on his challenge. It will be very interesting to “watch.”
Soden powerpoint which covers key points. See slide 7 for models with and without water vapor feedback.
The Big Red Dog is but one mutt in a litter of misconceptions that spawned most of the climate models. Sired by the mistaken notion that thermodynamic equilibrium within the system is achieved by radiative processes alone, it keeps barking at water vapor “feedback” as if it’s the postman in the sky bearing news of gloom, instead of a bald modelling fiat. That water vapor aloft is entirely the product of the surface-cooling mechanism of moist convection has been reduced to a parameterized afterthought in most model schemes. And that increased cloud formation and precipitation countervenes increases in net downward LW radiation is terra incognita, well beyond the Dog breeder’s tether.
On a planetary basis, nothing can increase the total thermalized energy in the system other than increased insolation. Everything else within the system is but an energy redistribution mechanism.
Re: John S. (#67),
I have to contest that statement. If that were true, then putting an extra blanket on the bed at night would not make you warmer. To a first order approximation increasing greenhouse gases in the atmosphere is identical to a decrease in the thermal conductivity of the atmosphere for outgoing energy while having no effect on incoming energy.
Re: DeWitt Payne (#69),
Because blankets inhibit convection, as well as adding materially to thermal capacity (insulation), the “extra blanket” analogy is totally misleading. Instead of inhibiting convection, the atmospheric “greenhouse” actually helps feed thermal energy into the moist convection that cools the surface. Meanwhile, the thermal capacity of the atmosphere, small to begin with when compared to that of the upper ocean and crust, is regulated primarily by highly variable water vapor. The e-folding “time constant” of the insulating effect of the atmosphere is very short, on the order of half a day, as is evidenced by the diurnal cycle. Once thermodynamic quasi-equilibrium for the entire system is established via all the mechanisms in play, changes in atmospheric capacity thus lead to negligible changes in surface temperatures at climatic time-scales. The effect of gross capacity changes would be seen mainly as damping of the diurnal cycle.
In any event, the total energy in the planetary system remains entirely dependent upon insolation. This is a fundamental matter of energy conservation, since GHGs produce not one calorie of it. Changes in albedo, of course, modulate the insolation that is thermalized. Climate models simply do not treat all these mechanisms realistically, as an integral system. And they ignore the conversion of thermal energy into mechanical work, which drives the winds and waves, with attendant losses to entropy.
P.S. What happened to your “kill list?”
Re: John S. (#72),
Sensible and latent heat transfer from the surface to the atmosphere by convection primarily occurs in the boundary layer that expands up to 2 km in the day and down to about 0.1 km at night. Above that level convection becomes much less important compared to radiation until essentially ceasing near the tropopause. Also, radiation is far more important in heat transfer in a closed room than you seem to think. Convection only becomes important if there’s a fan to move the air.
It’s “killfile” and it’s mental and informal and I’m becoming more forgetful with age. But thanks for reminding me. *Plonk* again.
Re: DeWitt Payne (#73),
Armwaving?
Re: jae (#75),
Skin temperature is about 35 C or 308 K. That means thermal radiation, ignoring emissivity, on the order of 510 W/m2 of skin area plus whatever latent and sensible heat is lost through respiration and the skin. The human resting metabolic rate, OTOH, is about 58 W/m2. According to this reference, a nude person sitting quietly neither gains nor loses heat at an external temperature of 31 C. Unless the room temperature is above 38 C, convection and conduction can only remove heat. That means that over 400 W/m2 must be coming from somewhere to keep the body temperature constant. Ignoring emissivity again, a surface at 31 C or 304 K radiates 480 W/m2 from the surrounding surfaces for a net loss of 30 W/m2. The other 28 W/m2 is lost primarily by latent heat from respiration and also from convection and conduction from the skin. Light clothing provides a closer radiative surface with some insulation so the external temperature can be lower, about 25 C. Numbers, not hand waving.
Re: DeWitt Payne (#73),
Troposphere general circulation is as it is just and simply because convection is NOT limited to the PBL!!!
Dynamics drives the vertical profile of the troposphere!
Re: Paolo M. (#78),
If that were solely the case, then a warm isothermal atmosphere would be stable because there would be no driving force for convection. However, a warm isothermal atmosphere that is capable of emission in the thermal IR is not stable to radiative energy transfer. The top of the atmosphere emits more energy than it receives and must cool. In the end, the temperature profile of a one dimensional radiative/convective model collapses back to about the observed temperature profile. In the opposite case of a cool isothermal atmosphere, convection forces the temperature profile towards the observed state. So both radiation and convection are important drivers of the observed vertical temperature profile.
Re: DeWitt Payne (#138 & #137),
your concept of Earth troposphere doesn’t resemble the true one.
The amount of energy (heat) in the troposphere goes up with height. You are confounded by the fact that temperature decreases, but you have to account for the decrease in pressure.
Air gets more heat with height, a very simple real fact.
Have you ever seen a cross section of potential temperature? I’ve found a plot of it here.
When you accept that temperature (potential) increases with height, you have to ask yourself what provides that extra heat to the upper troposphere.
It’s the latent heat of condensation.
Do you know what is the tool to carry it up there?
Deep convection!
If you don’t dismiss your idea that convection is important just in the PBL, you can’t understand how Earth’s atmosphere works.
The atmosphere engine is the equatorial convection. Everything has its origin there.
Have you ever wonder why the mean tropospheric lapse rate is pseudo-adiabatic?
And convection has the prime role in the cooling of the Earth surface, since it carries up all that energy (mass of water) collected by the trade winds in the tropic belt, not to mention the atmosphere general circulation.
Re: Paolo M. (#145),
Paolo I think you must have misunderstood the concept of ‘Potential temperature’, what you are suggesting runs counter to the kinetic theory of gases.
Re: Phil. (#147),
you are saying nothing.
If you wish I understand you, please, teach us how troposphere behave, if you can.
Re: Paolo M. (#150),
Rodrigo Cabellero’s (University College, Dublin) Physical Meteorology Lecture Notes is what I used to develop my understanding of how the atmosphere works. I’ve been too cheap to buy a textbook on this subject yet. Phil is correct, you do not appear to correctly understand the concept of potential temperature as used in meteorology. Try section 2.19 on page 39 of the link above.
Re: DeWitt Payne (#154),
Ignorance is not a guilt, it’s a state of our knowledge, maybe temporary.
I ignore so many subjects that I’m a patented ignorant.
But if someone, who never studied atmospheric physics, tell me that I misunderstood what potential temperature is,…well, let’s give up…
DeWitt, don’t worry about my ability to understand what P T is. You, who understand so well that concept, please teach us why potential temperature increases with height. I’m all ear and am looking forward to learn…
Or do you think that it doesn’t increase?
Are you, or Phil, able to answer this simple question?
Re: Paolo M. (#158),
Potential temperature is the temperature that a parcel of gas at altitude would have if it were adiabatically brought to a reference pressure (usually 1000 mbar). Just because a parcel of gas has a higher Potential temperature does not mean it has higher energy (your incorrect statement earlier Re: Paolo M. (#145)), the energy depends on the actual temperature.
Re: Phil. (#162),
I don’t know Atmospheric Physics in any detail, but it’d seem to me that a “potential” source of confusion is how potential energy is treated. The kinetic energy of the parcel of atmosphere at altitude might be rather low, but the total energy, which would include the gravitational potential energy might be much higher. The question is what energy is being talked about. I’d think Atmospheric Physics would have taken a position on the definition, but perhaps it’s not standardized worldwide.
Re: Phil. (#162),
here on Earth, if two parcels of dry air at the equatorial tropopause level have two different potential temperatures, be sure, I hope you not get upset, the parcel with the grater pT is the parcel with the greater energy. Have you ever imagined that?
Also, if a parcel of dry air with higher pT goes down from the tropopause to the surface, be sure…it will be warmer (grater T) than surrounding air.
Could we now go back to the problem core?
Which is the role of dynamics (convection) in the vertical profile of the troposphere?
Why the upper troposphere is warmer (pT) than the lower part?
Re: Paolo M. (#165),
All very fine and nice but it has nothing to do with what you said and which I took issue with!
Which is wrong and not addressed by your subsequent posts (also your reference to dry air is a cop-out in view of your earlier statement):
Where it appears that you are confusing ‘potential temperature’, which is for dry air, with ‘equivalent potential temperature’, which is for moist air and accounts for latent heat.
Re: Phil. (#167),
you are trying to make a lot of confusion and avoid the problem.
If I liked to talk about equivalent pT, I would have mention it.
What I liked to express is that upper troposphere gets energy from the latent heat of condensation, so I was forced to talk about pT and not epT. If you were an educated guy, you would have understood it quickly.
Sorry if I thought that some concepts were obvious, that I was referring to sensible heat, which is energy as you probably know, without adding that word. And, as you now probably know, to avoid problems with different levels of pressure, you have to consider potential temperature.
I think you are making a case because you are upset with me, aren’t you?
Re: DeWitt Payne (#169),
Ray Pierrehumbert (by the way, do you know him?), in the figure I invited you to look at, doesn’t fell the necessity to say what potential temperature is or how to compute it.
Why are you asking me?
Every educated guy knows what pT is and why is used in meteorology.
And yes, troposphere is generally stable thanks to deep convection!
And this is why convection in most areas is shallow, thanks to the equatorial deep convection!
Re: Paolo M. (#170),
Can you not see the contradiction in that statement? Stable in meteorological terms means no vertical convection. When the potential temperature increases with altitude, a parcel of air that is moved to a lower altitude will have a higher temperature and be more buoyant than the air around it. Similarly, if you move it up, it will be colder and less buoyant. The properties of water vapor, high heats of vaporization and fusion specifically, help establish the atmospheric temperature profile, but once established, there is little driving force for much additional convection in the atmosphere as a whole and radiative heat transfer becomes nearly as important as latent and sensible heat transfer at the surface (~70 W/m2 vs. ~100 W/m2) and rapidly more important at higher altitude. And that doesn’t include the contribution from direct absorption of incident sunlight by the atmosphere and clouds of about 60 W/m2 which also acts to lower the lapse rate/raise the potential temperature.
Obviously, heating of the surface by sunlight can create unstable temperature gradients. This can lead to a variety of behaviors from weak thermals to thunderstorms. But thunderstorms, while violent, only cover a small area at any given time and the vertical velocity of air in most thermals is not large compared to horizontal wind velocities.
Re: DeWitt Payne (#171),
there is no contradiction.
As you said often, convection is generally a shallow vertical displacement of air, linked to the planetary boundary layer, because the troposphere is generally stable.
Why the troposphere is stable?
Because the ITCZ convection releases most of the worldwide latent energy of condensation at a narrow band along the Equator. This warmed air, which you found high in the troposphere where the CB towers arrive, than moves poleward stabilising the troposphere, i.e. troposphere is warmer aloft
So yes, generally speaking troposhere is stable with the exception of ITCZ and the polar front belt.
And that thanks mostly to convection not radiation!
Dynamics counts.
Re: Paolo M. (#170),
Paolo M.:
January 7th, 2009 at 12:57 pm
No I’m trying to correct your mis-statements.
Which is strange because pT refers to dry air and epT refers to moist air.
Yes but that still doesn’t mean that energy increases with height in the troposphere.
No I have no reason to be upset with you, I’m correcting your mistakes.
Re: Phil. (#172),
your exaggerate obstinacy is going to make a person pigheaded (it is the traslation I found for the italian word “cocciuto”).
If your reaction to this claim of mine
is
that means you didn’t understand anything I wrote.
Perhaps it is my english or I’m not able to write clearly. So, before I try to state again, I invite you to read once more my words.
Moreover, since this little story is becoming really boring, if you give up now, that would be much appreciated by me and, I think, many others.
Re: Paolo M. (#165),
Potential temperature calculated how? The dry air potential temperature? Please cite your reference for this. The real atmosphere contains moisture and calculating the true potential temperature is not trivial. Btw, when the potential temperature increases with altitude, the atmosphere is stable, so there is no driving force for vertical convection.
Re: Phil. (#162),
I think it does in fact have higher energy. If a kg of gas has a higher temperature than another kg of gas at standard pressure, it would have higher kinetic energy. Inside the troposphere, though, it’s my understanding that potential temperature can be higher or lower than the surface at altitude depending on the local weather conditions. That’s why they take soundings. Above the tropopause, temperature, absolute and potential, increases because oxygen and ozone absorb short wavelength UV.
Re: DeWitt Payne (#73),
Apparently you have never seen anvil-top cumulonimbus reaching up to the tropopause. You’re free, of course, to believe anything you want, but if you can disprove the First and Second Laws of Thermodynamics, you can make a name for yourself beyond the blog world. Meanwhile, I’m grateful, once again, to be in your “killfile.”
Re: DeWitt Payne (#70),
If the atmospheric “heat pipe” (the rise of water vapor laden water) is responsible for a major fraction of the conductivity then decreasing the thermal conductivity of the atmosphere generally may not have much effect.
Think of doubling the insulation of a house (winter time conditions) while leaving the doors and windows open.
Of course the proportions are in doubt. However, the principle is right. The idea can be modeled by resistances in parallel. One resistor 100X the other. Doubling the 100X resistor to 200X will not change the overall resistance much.
So the deal is: if convection determines the atmospheric resistance to heat flow making the atmosphere less conductive without changing convection will not do much.
Re: M. Simon (#93), My take on the role of convection is as follows. In winter in Chicago we get layer clouds (when wind isn’t blowing) and stable atmosphere. As it warms up, convection cells form. At the hottest part of the year, huge convection cells are common. This sure looks to me like a negative feedback: hotter gives rise to increasing convection to higher and higher altitudes (an anvil thunderstorm can rise to 30,000ft +). It also seems to me that the big thunderstorms violate the simple calculations based on adiabatic equilibrium–they are out of equil.
Re: M. Simon (#93),
If there were a separate pathway for convective heat to be transferred directly to space your analogy to parallel resistance would be apt. There isn’t. Convective transfer occurs from the surface to the atmosphere. Most of that occurs in the first few kilometers. At 2 km altitude or 805 mbar, water vapor pressure is less than half that at the surface. At 8 km altitude (378 mbar), water vapor pressure is less than 1.5% of the surface pressure. All energy then leaves the system as radiation. The only open door is for radiation at wavelengths from about 8 to 12 micrometers where neither water vapor nor CO2 absorb strongly. Methane and ozone do absorb in this range, but there isn’t much methane and the ozone band is narrow. So a series circuit is more appropriate for the atmosphere as a whole and adding CO2 does increase the total resistance.
Convection in thunderstorms does bypass most of the troposphere. However, I have not seen any estimate of the amount of energy transferred by thunderstorms compared to the total incident solar energy or, more to the point, whether this transfer would increase significantly in a 2xCO2 world.
In the lower atmosphere convection is important. It becomes less important as altitude increases and becomes negligible at the tropopause on average. CO2 is an effective greenhouse gas because it has a very high molar absorption coefficient for radiation with wavelength near 15 micrometers. The brightness temperature of radiation emitted to space from the flat part of the CO2 band is about 220 K in the tropics (MODTRAN result) with the exception of the central peak emitted by warmer stratospheric CO2. Which, btw, is why increasing CO2 cools the stratosphere. 220 K is the temperature at about 200 mb pressure, or just below the tropopause. Increasing CO2 widens this band causing less radiation to be emitted at constant temperature
Using the Kiehl-Trenberth global annual energy balance, 102 W/m2 is transferred from the surface to the atmosphere by sensible and latent heat transfer. If the viscosity of the atmosphere were magically increased by many orders of magnitude so convection could not occur, the average surface temperature (isothermal body) would go from about 15 C (288 K) to 32 C (305 K), not 60 C.
Yes, but there is, obviously, some hitherto still unknown source that is powering the class A amplifier that drives the feedback in the earth’s climate./snark
Someone once, in here, attempted to tell me the effect is analogous to the resonance in a microwave oven. Apparently he missed the magnetron lecture. 🙂
Mark
Who else reckons that Steve’s reference to “Clifford, the Big Red Dog” derives from watching Nickelodeon with Bradley?
Is not the jet stream “convection” at a big scale?
Re: Craig Loehle (#78),
Technically, the jet-stream is “advection,” since the principal direction is horizontal, rather than vertical. But it is certainly a by-product of pressure differentials created by forced convection in response to the hydrostatic instability that develops under atmospheric radiative equilibrium.
P.S. Please let Geoff Sherrington know that I don’t have his e-mail adress. He needs to write me to discuss matters of mutual interest.
Re: John S. (#80), I did. Ball in his court.
Re: Craig Loehle (#82),
Re: John S. (#81),
Fixed, thank you both. For those interested, email is
sherro1@optusnet.com.au
Re: Geoff Sherrington (#105),
I’ll have to get back to you after I return from a trip.
Re: John S. (#81),
it’s a little obscure to me what you mean with
The high speed flow around 30°N and 30°S is ultimately caused by the equatorial deep convection and is a by-product of a rotating sphere. As you say, however, is not thecnically “convection”.
But, I don’t know if “forced convection” is technically the correct term for the equatorial convection.
Could you also elaborate about “the hydrostatic instability that develops under atmospheric radiative equilibrium”.
The other high speed flow you can find high in the troposphere is often associated to the polar front which, via the thermal wind equation, has its origin in the strong surface temperature gradient across the polar front itself. In this case, convection is not required.
Re: Paolo M. (#91),
My aim was not an exposition of the dynamics of the jet-stream per se, in which planetary rotation certainly plays a major role, but merely an acknowledgement that advection is connected to pressure differentials developed by convection.
As for convection forced by radiative equilibrium in an absorptive gravitationally-bound atmosphere, that is a long-known analytic result going back to Gold and Emden, and discussed by Milne, in the early part of the last century. Purely radiative equilibrium leads to grossly superadiabatic lapse rates, i.e., a hydrostatically unstable, top-heavy atmosphere. Thus violent convection is physically necessary to establish thermodynamic equilibrium with a sustainable adiabatic on a dry planet. On a waterous planet, condensation aloft relases the latent heat of evaporation and helps flatten the lapse rate, thereby reducing the required strength of convection of sensible heat.
Hope this clarifies what I was referring to.
Re: John S. (#103),
sorry, I didn’t understand you was referring to that non-exsistent, academic planet.
Re: Paolo M. (#108),
Dry planets do exist, e.g., Mars. And on Earth, convection remains quite indispensible in establishing thermodynamic equilibrium. But let’s not belabor this side-issue.
Re: John S. (#109),
we all live on Earth and Earth is what we are dealing with, aren’t we?
I didn’t understand your remark on convection trying to estabilish thermodynamic equilibrium, since it was my point in comment #78…
And convection is not a side-issue as it’s a crucial point to understand Earth’s behaviour and someone here (not you) hasn’t realized yet.
I wonder why they make convection ovens?
Or why we have HV/AC forced air system in our homes ?
old construction worker,
“Or why we have HV/AC forced air system in our homes ?”
limited time??
I used to live in an old house with ‘gravity heating’. There was no blower; it used the ‘hot air rises/cold air falls’ principle to cause air to circulate. It was very quiet but did not heat evenly.
Steve McI,
Not to be boring but could you fix your code?
#88. That should be dim(instr)[1] not dim(instr)[3]. I fiddled a bit with the code and must have forgotten to update a segment. Try it now.
Opti – to see – from optical.
mal – bad or badly.
So the climate “science” definition of optimal = to see badly. i.e. we just see what we want to see.
the rise of water vapor laden water
should be: the rise of water vapor laden air
Steve, you’ve made it worse!
It’s all length
Of optical path tau
And Simon auld lang syne.
===================
Steve, could you comment on Hu’s #37?
#95. John A – you trying to do this without the packages ncdf, fields.
Re: Steve McIntyre (#100),
Holy crap! It all works! It’s so simple that even a climate scientist can run it!
Steve, I take back all those nasty things I said about you.
I’ve edited the code to add some vital information for R-idiots like me.
Re: John A (#104), I tried these now that I’ve finally installed R on my machine, and the UAH script produces a pretty graph. The RSS script fails like this:
> ###CALCULATE TRENDS
> Trend= array(NA,dim=c( 5,18))
> for (k in 1:5) {
+ year=c(time(zonal[[k]]))
+ temp=(year>=1979)
+ for (j in 1:18) {
+ fm= try(lm(zonal[[k]][temp,j]~year[temp]))
+ if (!(class(fm)==”try-error”)) Trend[k,j]=fm$coef[2]
+ }
+ }
Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, …) :
0 (non-NA) cases
Error in lm.fit(x, y, offset = offset, singular.ok = singular.ok, …) :
0 (non-NA) cases
>
> ##PLOT CONTOUR MAP
> #library(fields)
> hpa=c(1000,740,466,220,75)
> range(Trend,na.rm=T) #-0.08826997 0.05186060
[1] -0.07743676 0.05186060
> breaks0=c(-0.095,seq(-.055,.055,.01),.095);n=length(breaks0)
> filled.contour(x=lat,y=1000-hpa,z=t(Trend),levels=breaks0,col=tim.colors(n-1),main=”RSS/CRU Trends”,
+ plot.axes = { axis(1, seq(-60, 60, by = 20))
+ axis(2, at=1000-hpa,labels=as.character(hpa)) } )
Error in filled.contour(x = lat, y = 1000 – hpa, z = t(Trend), levels = breaks0, :
dimension mismatch
I fixed the “d:” paths and the extra space in v <- open.ncdf(“temp.dat”). Everything that precedes the “CALCULATE TRENDS” section completes without error.
Re: Not sure (#186),
I am a R newbie (Windows v2.8.1) and also got the “dimension mismatch” error in filled.contour. I may be wrong, but I figured this was because “lat” has dimension 72 whilst the corresponding dimension of “Trend” (and of its transpose, “t(Trend)”) is only 18. I solved the problem by adding this line before the PLOT CONTOUR MAP line:
zone=seq(-85,85, by = 10)
which generates 18 zonal latitude values, and then changing “x=lat” to “x=zone” in the filled.contour line.
The correct graph then seems to be plotted. But if my theory is correct then Steve and others shouldn’t have got the script to work OK without making this sort of change. So I am probably missing something.
Steve: I’m trying to make these scripts turnkey, but they are done quickly. Sometimes I edit things in the console and there may be a loose end on my script as posted. LEt me know and I can usually figure it out pretty quickly.
Re: Nic L (#189), Thanks Nic, that did the trick. I get a graph that looks a lot like the one in the post. I still get the two “Error in lm.fit…” messages, though. I also appear to get different results from what Steve expects in this line:
> dim(instr)# [1] 72 36 1899
[1] 72 36 1907
Steve: Ever consider using source control?
(deg C/year). I didn’t calculate the conversion to 2xCO2 tho this should be straightforward on usual assumptions.
Here’s GRL’s submission URL for Spencer and RomanM to send their critiques:
http://grl-submit.agu.org
Speaking of uncertainty, it would be helpful if Steve ammends his post to point out that the temperature trends from the satellite “channels”, and therefore the trends in his graphs, do not represent temperature trends at those specific altitudes.
(And that should be “RAOBCORE” not “RAOBCARB” a few posts ago)
Re: cce (#102),
Why should submission be required to do the job that should have been done by the peer review before the paper was published? I guess that it must have met the scientific standards of the pro-AGW consensus. Please tell me honestly, did the omission of uncertainty considerations not bother you? Or does the the hockey stick Mann-tra apply: It doesn’t matter because the result is “correct”…
Re: RomanM (#107),
RomanM,
I missed your critique of jae’s Spencer link. Please repost it so we can verify that you care about that which you speak.
Dessler et al gave the entire range of their results. That is, 0.94 to 2.69 W/m^2/K. Models predict 1.6 – 2 W/m^2. As they state in the paper, the specific humidity and atmospheric temperature from the AIRS satellite are accurate to 10% and 1 degree respectively, and they compared December-January-February averages. Given this, do you think “uncertainty considerations” are going to allow you to fish a small or negative number out of those equations?
[I’m also patiently waiting for an update to this post describing how the satellite channels do not represent the temperatures at those specific altitudes.]
Re: cce (#112),
I hadn’t realized that you had appointed me as sheriff in charge of policing all papers in climate science. My post was a response to your established method of throwing out references to entire papers (in this particular case, of somewhat dubious scientific merit) instead of pointing out and arguing the specifics items in the papers which may be relevant relevant to the situation. Why don’t you tell us where, in your opinion, Spencer goes wrong? That’s how scientific discussions take place.
From your last sentence, I surmise that your answer to my previous question: “Please tell me honestly, did the omission of uncertainty considerations not bother you?” is as I suspected: “It doesn’t matter because the result is ‘correct'”. Maybe you don’t quite understand what is involved here. The authors use Equation 1 to calculate the estimated feedback parameters in Table 1. If you look at the equation, you will notice that there is a delta-T in the denominator of one of the terms. The authors explain that “Each row lists the feedback calculated using delta-q
and delta-Ts between January 2008 and one previous January.” Calculations involving terms with errors in the denominator of expressions have a notoriously strong (non-linear) impact on the overall result, particularly when those values themselves are close to zero. Your “accurate to 10% and 1 degree” assessment of variability is extremely naïve and an estimate of the variability of the final result is absolutely necessary to indicate how uncertain that result is. With only five years worth of estimates (which are all calculated relative to the same base year ensuring a lack of independence between those estimates) and with the reliance on all of the other tacit assumptions inherent in the paper, who knows, we might possibly be able to “fish out” anything from a rubber tire to a whale.
Actually Steve, when I look at the code output I see:
Since this is probably a fragment you were saving on your laptop for other reasons but is not required to produce the graphic, perhaps it could be remarked out?
A little off topic, anyone actually validated the Mauna Loa, Hawaii CO2 readings?
If I measure close to the same place will I get the same ppm number that IPCC and the radical environmentalist uses?
Instead of being either a positive or negative feedback wouldn’t it be safe to assume, given the history of the planet, that water vapour is a moderating feedback?
Feedback has to be either positive or negative, Jeff, as that is the sign of the value that gets fed back.
Mark
A laymen’s account of the development of AGW “fingerprint” criteria can be found in William K. Steven’s “The Change in The Weather: People, Weather and the Science of Climate” (1999), chapter 13. It mostly focuses on the career of Ben Santer. Steven’s was a NYTimers reporter, and the chapter does conclude with a note of uncertainty.
Nice collection of tropical tropopause reprints from JPL ( link )
Re #119 Second try for the link
Re #120 Last try
http://mls.jpl.nasa.gov/research/ttl.php
Mark T: January 4th, 2009 at 5:23 pm
Yes…
Wasn’t that my point?
Water vapour could provide a positive or negative feedback depending upon what it is feeding back on. A moderating influence that prevents extremes in either direction.
Re: Jeff Norman (#122),
I’ve never heard of any definition of feedback that does such a thing. Stable systems with feedback, i.e., no poles outside of the unit circle, always result in a bounded output given a bounded input, without knowledge of where the system resides. The feedback can be either positive or negative (or zero, which is trivial), but not both at the same time.
Mark
I guess that would be usually prevents extremes in either direction given the periods of glaciation. But who knows, maybe locking up all that water in ice and drying out the atmosphere is a feedback against something worse.
I think the term used here should be simply and purely, “negative feedback”; – that is to say, feedback producing homeostasis in the system, as in the standard case of the centripetal governor on a Watt steam-engine, that opens or closes valves depending on the speed it is rotating at, thus ensuring that the engine is kept from running too fast or too slowly. This is precisely what “preventing extremes in either direction” means. One should therefore be a little more charitable to Jeff Norman. His point was clear. Positive feedback, on the other hand, simply acts to alter the state of the system by accentuating the disturbances producing the feedback signal. The expansion of the poles is sometimes held to result in a positive feedback causing runaway ice ages via the increasingly greater reflective effect on the growing ice-cover. Some of the scarier scenarios of climate change postulate a runaway positive feedback loop via methane released from warming tundra. Reading the posts on this site is a useful corrective to the scare-mongers.
Re: Kimberley Cornish (#125),
You state that as if I were attacking him, I was not. And no, his meaning was not clear, and his follow-up was equally confusing (either/or?). There is a reason standard terminology exists, it allows people to communicate clearly.
Mark
Well I like the idea that oceans and watervapor is a thermostat that regulates the the Earths climate around Equator to around aprox 28-30 deg C. If the temperature is below then the oceans and watervapor tend to be more positive feedbacks. And if it gets above they tend to be more negative feedbacks.
If it is correct that the Earths climate and global avg temperature actually/theoretical should have been more than +60 deg C.
And that the reason for it beeing only 14-15 deg C is that huge amounts of “solar” energy is vented out of Equator vertically/horizontally it means that we on the total have a huge negative feedback in Earths climate.
Because like when you boil water it will not get above 100 deg C, without increasing the airpressure, And that the same applies to the Equtors Oceans and its rich watervapor atmosphere that even if the sun shines 24 hours a day it still would not go above 28-30 deg C?
http://www.drroyspencer.com/research-articles/satellite-and-climate-model-evidence/
“The resulting picture that emerges is of an IN-sensitive climate system, dominated by negative feedback. And it appears that the reason why most climate models are instead VERY sensitive is due to the illusion of a sensitive climate system that can arise when one is not careful about the physical interpretation of how clouds operate in terms of cause and effect (forcing and feedback).”
To jon-Anders: Nope. In BOTH cases it is negative feedback.
To Mark T: I didn’t actually say that you were attacking him. He was simply wrong about the distinction between positive and negative feedback, but it was not materially significant to the point he was trying to make. What you wrote was indeed quite correct, but it was not, as it were, necessary to correct him. I wouldn’t describe what you did as an “attack” at all. It is just that in disputation sometimes it is reasonably clear what someone is trying to say, even if they express it badly, or even (by the letter of the law) falsely. Sometimes a flyswatter should be used kindly.
Re: Kimberley Cornish (#129),
Forgive me for noticing but you did the same thing with Jon, i.e., you corrected an obvious (to you and I) error he made. I disagree however, people need to state things properly else this site earns a reputation that it does not deserve. I was merely pointing out his error as you did with Jon. Whether he chooses to use that opportunity as education or insult is up to him, but I only meant the former.
Re: Jon-Anders Grannes (#131),
You don’t even need “more or less” to get the effect you’re talking about, Jon, particularly if the “inputs” are oscillating. Also, as Kimberly implied, positive and negative do not necessarily result in up or down, either can result in both effects.
In general, there are many misconceptions about what feedback does and how it works. On the surface it is a simple subject, so this is no surprise.
Mark
“It has recently been recognised that it can be cast as a multiple regression problem with respect to generalised least squares”
Go Gauss!
reply 129
Okay I agree.
We are acyually taliking about more or less negative feedback and that around 28-30 deg C it starts too be more over the Equator ocean ?
There is also a suggested “AGW fingerprint” in Kaufmann & Stern: Evidence for human influence on climate from hemispheric temperature relations, Nature, 1997 (see also here). That “fingerprint” is the south-to-north (Granger) causal order of the surface temperature. I know that the “AGW fingerprint” “nature” of the finding was disputed by Triacca (“This paper was submitted to Nature but rejected.”; see also here; ), but I wonder if anyone is aware what is the status of the south-to-north causal order finding? Is it still true? Came to my mind when I saw this:
Re: Jean S (#132),
Interesting article,
Re: Jean S (#132), According to several papers I have seen, there is a historic (over many tens of thousands of years) see-saw effect with north and south poles alternating warmer and colder over decadal scales. Thus your figure is only to be expected and has nothing to do with AGW (NOT a fingerprint).
Re: Craig Loehle (#139),
Let me try with another figure. In Kaufmann paper data from 1865 to 1994 was first used to show that SH temperature Granger causes NH temperature. Mann’s lowpassmin was handy, so I used it to smooth HadCRU NH and SH to decadal:
Now, it seems that NH should get colder soon ? Next step of Kaufmann paper is to show that this causal order is a fingerprint of anthropogenic activities,
..but it seems to me that something has changed. I guess this idea of retarded NH temperature came from 1970-1994 data ?
Re: UC (#144),
That’s a phrase I’d never have expected you to use! 😉 Sorry, couldn’t help but chuckle. Good thing I didn’t have any coffee in my mouth at the time.
Mark
Re: Mark T. (#151),
I wanted to exaggerate the difference a bit 🙂 Nice to be a part of the denialist machine, btw.
Re: Craig Loehle (#139), Re: Bill Illis (#146),
The supposed fingerprint in K&S is the south-to-north causal order not the difference per se between south and north. The Granger causality there means that you can use the past temperature from south and north to predict the northern temperature better than to predict the northern temperature using only the past northern temperatures. The causal order means that this does not hold the other way around (predicting southern temperatures using both northern and southern data is equal to predicting southern temperature using past southern temperatures alone).
Thus my inquiry is if anyone has tested if this south-to-north causal order still holds according to the most recent data. If it does, as far as I understand, this should mean that the north should follow (at least somewhat) what is happening in the south. That was my figure about.
Re: UC (#144),
This is the thing that was shown to be false by Triacca. That is, the causal order is not a “AGW fingerprint” of the causal order (or at least, K&S’s analysis does not show it). However, Triacca does not dispute the causal order finding itself.
Re: Jean S (#132),
The difference in the northern and southern hemisphere can be explained by the different natural ocean cycles affecting each – the ENSO, the AMO and its southern counterpart in the south Atlantic.
When this is accounted for, the northern hemisphere, the southern hemisphere and the tropics are all warming at almost exactly the same (low) rate (between 0.09C per decade to 0.11C per decade – half the rate predicted in the models).
As Craig Loehle mentioned, the multidecadal ocean cycles (including the ENSO which is not multidecadal) are the following:
Re: Jean S (#132),
In discussion on the related news of Christy’s report of three decades UAH temp anomaly at Andrew Watts WUWT
Commenter John Cooper cited a publication in 1997 by S. Fred Singer, which may explain this zonal temp anomaly. The claim using data from Christy and Spencer is that NH airline traffic generates tropospheric increased amounts of water vapor, increasing the total amount and therefore radiative efficiency of cirrus clouds in the atmosphere. Surely many have read this before: increasing low clouds reflect incoming UV back into outer space; increasing high cirrus clouds reflect IR back to earth.
Here is Cooper’s cite (from above link) in full:
reply 132
This shows that the ventilation of energy from the Equator the last years is ventilating more to the Norht and less to the South?
#135,
I’d guess the point is that
1. Statistically oriented paper that shows the human influence in global temperature record is published in Nature
2. Criticism of statistics in 1. is submitted to Nature, but rejected. Déjà vu.
3. In addition, the fingerprint mentioned in 1. seems to be missing.
Mark’s correct. I should have said:
Instead of being either a positive or negative forcing wouldn’t it be safe to assume, given the history of the planet, that water vapour is a moderating forcing? But then water vapour isn’t a forcing but a feedback.
But then the AR4 WG1 SPM says:
And then the AR4 WG1 Chapter 8, Box 8.1 says:
And of course in IPCC parlance positive feedback means warming. So to keep with the program I used negative feedback to mean cooling.
I did not claim that this would make sense to a control engineer, but then little of this does.
Re: Jeff Norman (#140),
Yup. Some in the climate world seem to understand the difference, but not all.
Yes, I understand what you’re getting at, and I think these definitions are where the misconceptions lie.
I’m actually a signal processing engineer, but control problems are part of the business. My guess would be Kimberly is, however, a controls engineer (or something more closely related). You are correct, the way these things are often described can be best described as… painful. You should go find the threads where Jean S discovered how Mann centers his proxies (RegEM threads). Dumbfounded is the best way to describe the reaction.
Mark
Re: Jeff Norman (#140),
I understood your original point to be postulating water vapour to be a negative feedback, albeit with some confusion as to terminology (“moderating” etc.).
However, I can’t agree with this statement:
IPCC AR4 8.6.1 (p. 629) says:
So in “IPCC parlance”, positive feedback amplifies the effect of a given forcing (whether that effect is warming or cooling), a usage of the term compatible with that in other fields as far as I can see.
I guess I’m not the only one.
A Glimpse Inside the Global Warming Controversy
Mark,
I’ve been here since the beginning and have seen all the threads. I just usually have the good sense not to say anything.
I hope this is not seriously out of line, but let me ask the question. It seems to me that McKitrick & Michaels (and de Laat) have convincingly demonstrated that up to 50% of the warming seen in the major indexes could be spurious (i.e., related to economic development rather than CO2, see here).
If you adjusted/reduced the surface warming by 25% or 50%, would you then see the predicted “hot spot” in the tropical troposphere? That is, comparing the adjusted surface temperature to the measured tropical troposphere temperatures, would you then see the expected “amplification” of the surface temperature?
Wow. Sad, actually.
Mark
lowpassmin result shows that decadal NH-SH difference is now largest ever. You’ll get different result if you take difference first and smooth then (by 0.1 C) but still largest diff. Quit a non-linear smoother.
Jean S, you think this is the data Kaufmann used ?
Re: UC (#155),
Yes, updated with one more year.
OK, am I missing something? I decided to reproduce Kaufman and Sterns Model 1 (temperature only) analysis. Not having their NH and SH data, I went to HadCrut and used that to fit the regressions in the paper for Granger Causality. The R program to do that is
The results from R were
The data I used gave:
South “causes” North: F = 0.7847, p-value = 0.5373
North “causes” South: F = 0.7984, p-value = 0.5285
No significant results here!
These are a lot different from the values obtained in the paper:
South “causes” North: F = 3.19, p-value = 0.016
North “causes” South: F = 0.42, p-value = 0.79
Exactly what did their data look like?
Re: RomanM (#157),
Thanks! Yes, I know I should do it myself, but does the data linked in #155 give similar results (2nd column is NH and 3rd is SH)?
Re: Jean S (#159),
It wasn’t too difficult to adapt the previous program to do it. Assuming no errors (after dinner and wine)
Despite the fact that year 1994 is missing, you would not expect that the p-value for South “causes” North would increase from .016 to .2796. Are you sure this is the data they used?
Re: RomanM (#160),
Well, their reference is
The data in #155 comes closest to that from what I’ve found. Contact information for Kaufmann is here and for Stern here if someone wants to request the actual data used.
Now this is getting interesting. Not only K&S’s conclusion was premature, it also seems the effect they found is non-existent.
Re: Jean S (#166),
Rather than asking for the data used in the Kaufman and Stern paper, I decide to search around and see what I could find. In all, I located seven data sets for Northern and Southern hemisphere temperatures (there could be more) which I thought would be relevant or interesting to look at in this context. All of them have their origins in CRU.
Briefly, using the names I have given them:
For this data, I ran the same analysis as for their model 1: Regress the temperatures on time and the first four lags of the series. Then to look for “cause”, include the first four lags of the series from the opposite hemisphere. An F-statistic and p-value for the added variables were calculated – the same statistics used in the in the paper. The time frame in the paper was 1865 to 1994. I also ran several other time frames to see what the results would look like. (They would be easier to read in a nice table format):
It looks to me that the dataset used in the paper was (close to) tavenh.dat since their results were:
South “causes” North: F = 3.19, p-value = 0.016 (mine: 3.1321 and .0172)
North “causes” South: F = 0.42, p-value = 0.79 (mine: 0.4113 and .8003)
Curiously Jones and HadNH were considerably different for the same period.
Looking at the data from 1900 to the present, it is interesting to note that the “causal effect” is not particularly present in any of the data sets except the smoothed HadsmNH. However, as in other filtering situations in regression, it is pretty much going to be the case that smoothing will create the desired effect even if it isn’t there to start with.
Re: RomanM (#160),
I get the same values with Matlab,
(y’*(P-P0)*y/4)/(y’*(I-P)*y/119)
ans =
1.28540117893769
But something worries me here, we have a model temp= bias + trend + AR(4) , any room for CO2 ? Secondly, why is
rank(I-P)
ans =
120
And then, we have y= Xb+ e where X is a bit stochastic, but I guess that’s OK in this case, need to read Yule-Walker chapter again 😉
Re: UC (#180),
I don’t know why the degrees of freedom would be 120 in the Jones dataset case. The dateset ends in 1993 so there are 129 observations in the sequence. If you subtract df’s for the 10 estimated parameters (constant term, trend slope, and 2 sets of 4 AR terms) you get 119, as advertised. Do you have an extra line of data or one less parameter somehow (unlikely since we have decimal agrement to so many places). Maybe its a “bug” in Matlab. 😉
What do you mean by “any room for CO2”?
Re: RomanM (#181),
Rounding problem ? rank.m provides just an estimate of the true rank..
I=eye(129);
P=Xb*pinv(Xb);
P_=Xb*inv(Xb’*Xb)*Xb’;
rank(I-P)
%ans =
%
% 120
rank(I-P_)
%ans =
%
% 119
% but
max(max(abs(P-P_)))
%ans =
%
% 2.0178e-014
Re: RomanM (#181),
Model 3. in Box includes trend and CO2 . Model 1. doesn’t have CO2, but is there really anything left for CO2 to explain ?
Some Model 1 realizations, assuming LS-solution is correct:
Re: RomanM (#157),
I’ll be happy to send you our data and programs (in RATS) if you e-mail me. We used the Jones temperature series. OTOH I don’t think the Nature paper was our best work, just the first one we published. There are links to all the later papers at my website.
Re: David Stern (#191),
Thank you for your kind offer. It is always a lot easier to understand how someone has done an analysis when they share their materials and answer questions about what (and why) optional choices were made in their procedures. I will drop you an e-mail shortly.
Current soundings in the tropical atmosphere, including potential temperature, can be found here:
http://weather.uwyo.edu/upperair/sounding.html
Indonesian stations or the South Pacific islands are reasonably representative of areas of moist ascent. Hilo, Hawaii is about the best at this website for the downleg of the Hadley-Walker circulation though I wish Galapagos was available.
David, how do you manoeuvre to the digital data? I got blank pages.
Hmmm- the University of Wyoming website seems to be having problems at the moment on the particular upper-air page.
Normally what appears are regional maps. A user clicks on the station of interest, causing a menu to appear. The menu gives the user a choice of data presentations (numerical, plotted, date range, etc).
I assume the website will be back shortly.
Re#175 The sounding website is now OK. The menu may be hidden when the page opens – simply scroll up.
Some of the moving parts of the tropical troposphere become visible when one plots and plays with this tropical sounding data. At times I wonder if all the upper troposphere processes are identified and understood, especially in the convective regions.
A good source for physical interpretations and data on the tropical upper troposphere is this collection of Andrew Gettelman’s papers. He writes well and reading them is relatively easy.
Here’s a sounding plot for Menado, in the thunderstorm region of the western Pacific:
It shows the change in temperature, potential temperature (a parcel brought to 1000mb) and equivalent potential temperature (includes a release of any latent heat in the parcel).
The wiggles in the red line around 5km probably represent mixing of dry air into the layer. Above about 10km the air is quite dry – there is little latent heat in a parcel so the parcel’s potential temperature and equivalent potential temperature are about the same.
And here is the current sounding for Hilo, Hawaii US:
(I marked the midlevels as dry descent but actually Hilo is experiencing some moister midlatitude air intrusion at the moment. It’s not a perfect example.)
The inversion, which is common, is distinct. Below the inversion lies moist near-ocean air. Above the inversion is the dry, radiationally-cooling and sinking air originally lifted high into the atmosphere in thunderstorms thousands of miles to the west.
The equivalent potential energy (known as theta e) increases suddenly below the inversion, indicating the high moisture content.
Regions like the air above Hilo, where radiational cooling is important and water content is very low, are where I’d expect to see a distinct effect from increased CO2, a fingerprint unlike any created by solar changes.
Re: David Smith (#183),
the two soundings you plotted seem to be a good example of the role of convection and radiation in the final result far away from the Equator.
All the energy released at the ITCZ is only partly radiated back to the outer space. Over Hilo, as well over the whole globe, upper troposphere is still warmer then the surface.
Troposphere is generally stable because of ITCZ deep convection.
Another consequence of this tropospheric behaviour is that an eventual increase of water content in the lower levels of the atmosphere (due to CO2 IR absorption, for istance) would enhance the ITCZ activity whereas the region of dry descending air would expand, leading to a less water content in the medium-high troposphere outside the ITCZ region and to a widening and/or thickening of the stratocumulus clouds layer, not to mention the Iris Effect.
yep, shouldn’t use rank with default tolerance ,
>> rank(I-P,10^-10)
ans =
119
>> rank(I-P_,10^-10)
ans =
119
Sorry about that 😉
One Trackback
[…] Gavin and the Big Red Dog « Climate Audit […]