Thursday, February 11, 2010

Hot Water and Air

Georg Hoffmann, writing in German on Prima Klima, has an excellent post on Susan Solomon's new paper on stratospheric water vapor. Eli with permission is reproducing it in translation here. It goes down a treat with Eli's post on atmospheric methane oxidation but goes far beyond that and offers great insight into stratospheric cooling.


How water vapor in the stratosphere can occasionally halt global warming!

A paper
has appeared in a recent January issue of Science written by Susan Solomon, the former IPCC Working Group I Vice-Chair, and her colleagues, discussing the influence stratospheric water vapor. There have been a few very useful commentaries on this paper (here and here) and I will postpone until later observations on where and why it was often, but not very usefully mistreated. As usual, the latter variety of comments was in the vast majority.

OK: Water vapor in the atmosphere. Everyone is aware that it cools as you rise higher in the atmosphere and as a consequence it has to become dryer. If there was no water on earth but all the other greenhouse gases remained in the atmosphere, we would be faced with a temperature gradient of about ~ 10 °C/km (this is the so called dry adiabatic lapse rate), but as it rises water vapor condenses, liberating heat, which reduces the gradient to the observed ~ 6.5 °C/km, (the saturated adiabatic lapse rate). I'm not quite sure what exactly defines the troposphere, but a key feature is presence of the saturated adiabatic lapse rate that that holds up to an altitude of about 10-12 km.

Figure 1: Radiosonde profiles of water vapor concentrations measured with an IR laser absorption spectrometer by Georges Durry and co-workers at the University of Reims,.

But what really happens with the water vapor above the troposphere? Is there any water at all, and if so how much and is it of any significance? Colleagues from the University of Reims have periodically flown balloons carrying a laser spectrometer (you can, of course, also do this with conventional hygrometers) to find out about this. Be careful when you look at Figure 1, the concentration axis is logarithmic. The concentration of water vapor decreases from about 10,000 ppm at the ground to about 3 ppm at 15 km altitude, almost nothing. But strangely enough, the water vapor content then increases again.

Figure 2: Schematic of the processes giving rise to exchange between stratosphere and troposphere. Taken from the Heidelberg dissertation of Frank Weidner.

It turns out that water vapor in the stratosphere has two sources. The first, of course, is transport through the tropopause, the boundary layer between the troposphere and stratosphere, (see Figure 2). This sounds trivial, but it is not. The transport takes place mainly in the tropics, either through abrupt short-term events driven by very intense convection, sometimes shooting up to over 15km, or via so-called tropopause folds where neighboring stratospheric and tropospheric air masses exchange places and become intimately entwined. Figure 3 shows an example of such a lowering of the tropopause along a frontal zone in the North Atlantic. In the image where very dry stratospheric air is mixed into the troposphere the figure appears suddenly dark which corresponds to dry on the satellite false color scale. It is generally assumed that most of the water vapor found in the stratosphere passes through the tropical tropopause which is the coldest point of the troposphere, into the stratosphere.

Figure 3: Tropopause folding. The extremely dry stratospheric air is mixed into the troposphere and leads to this "black hole" in the water vapor concentration. Meteosat image from ZAMG.

The second source of water vapor is the photochemical decomposition of methane in the stratosphere which only comes into play at high altitudes, explaining the increasing concentration of water vapor at the left of Figure 1. If there were no methane oxidation, water vapor concentration in the stratosphere would remain almost constant. Whoever wants to know more about the chemistry of methane in the stratosphere, should read the post by the Godfather of all climate blogging, Eli Rabett, for details. Caution! Not for the skeptics: It’s very informative.

It is estimated the contribution of methane to stratospheric water vapor is of the order of 30% (see Figure 1) with an increase of between 4 to 5.5 ppm from the lower to the upper stratosphere. In most of the 20th century atmospheric methane concentrations have risen with expanding livestock production, growing waste dumps and increased rice cultivation. It is widely expected that the methane concentrations will continue to rise. Therefore we must expect a long-term increase in stratospheric water vapor, assuming that everything else remains the same.

But will everything else remain the same? There could be more or fewer folds at the tropopause, more or less intense convection in the tropics, but above all, could the "cold trap" warm. The "cold trap" is the extremely cold zone in the tropical tropopause through which most of water vapor has to pass to reach the stratosphere, warm. In short, we need long-term observations from the stratosphere to say something reliable. In their Science paper Susan Solomon and colleagues have combined a couple of data sets. Figure 4 shows the radiosonde water vapor data set from Boulder, which to my knowledge is the longest such record, together with various satellite records. While the Boulder, Colorado measurements have been taken at a single location this is not thought to be a big problem. As far as water vapor is concerned the stratosphere is well mixed. Thus, we can compare the point measurements from Colorado with all sorts of geographically distributed measurements from satellite observations (HALOE and SAGE). The bottom line from all this: Stratospheric water vapor increased for twenty years but since 2000 has decreased a bit.

Figure 4: Different observational datasets of stratospheric water vapor. The radiosonde data from Boulder, with two satellite data sets.

So what? One can ask this, because, God knows, the change in stratospheric water vapor is so tiny, on the order of 1 ppm. Yes, but it makes a difference. Water vapor absorbs and emits in the infrared, just like the classic greenhouse gas CO2, CH4 and N2O and a change of 20% in stratospheric water vapor as seen in Figure 4 may well have an important influence. Unfortunately, it is really complicated to address the issue of “greenhouse gases and their effect on the stratosphere ". It is a hell of technical details. When greenhouse gases increase the first thing that happens is that the amount of infrared radiation from the surface reaching the stratosphere decreases. Thus, the "primary" effect of increasing greenhouse gas is to cool the lower stratosphere. In addition, since greenhouse gases in the stratosphere also radiate, increasing their concentration more effectively cools the levels they are at by radiative cooling. This stratospheric cooling is very fast, since temperature gradients in the stratosphere are determined almost entirely by radiative transport.

But there is a second long-term effect of rising greenhouse gas concentrations. Right! Infrared radiation in the troposphere is increasingly converted into heat, which warms the troposphere, although this takes quite a long time to reach a new radiative balance (oceans, melting glaciers, sea ice, the whole rigmarole takes its time). Thus, after the greenhouse gases rapidly cool the stratosphere by shielding the infrared radiation from the bottom, they warm up the troposphere on a longer time scale (~ 10 years). This second step increases the incoming heat flux to the lower stratosphere, somewhat mitigating the first cooling effect. How much is not trivial to figure out. For that one need to a detailed line-by-line radiative transfer model.

But all this is unfortunately not everything. What I have told in the previous section on greenhouse gases holds only for CO2 where the main absorption band at 15 microns is saturated over a distance of a few meters. Because of this, in the lower stratosphere infrared radiation in the CO2 bands can only come from the cold upper troposphere. In that case, the above-mentioned shielding mechanism is how the greenhouse gases work. With methane and N2O, things are different. [ER-Because the concentrations of these two molecules are much lower] IR radiation in the regions where they absorb and emit passes directly from the warm lower troposphere to the lower stratosphere. An increased concentration of these gases has virtually no effect on the net radiative balance of the stratosphere. On the one hand, there is more IR from lower layers to absorb, on the other hand, there are more radiators.

Figure 5: Observed and calculated warming of the last thirty years. A simple climate model was used driven by the accepted history of changes in greenhouse gas concentrations (dotted line), then with the modifications taken from the stratospheric water vapor mixing ratios for the last 8 years (red) and then with the entire observational record of stratospheric water vapor since the beginning of the 80s (blue). The observations of surface temperatures (NCDC, GISS and CRU) are indicated in green.

And what about the water vapor in the stratosphere? The stratospheric water vapor will contribute as an effective radiator to increase radiative cooling of the stratosphere and warm the troposphere. Water in the atmosphere is not well mixed (obviously, see Figure 1) and each water molecule in the high troposphere or lower stratosphere, respectively, is a very effective infrared absorber / emitter. Solomon calculates a radiative forcing from stratospheric water vapor of about +0.25 W/m2 in the period 1980-2000 and ca -0.1W/m2 since. For comparison, the increase in CO2 over the same period accounts for a forcing of 0.36W/m2 for 1980-2000 and 0.26/m2 since. Solomon then throws this radiative forcing into a simple climate model to quantify its impact arriving at a stratospheric water vapor contribution of 30% to global warming since the early 80s and a contribution to the flat course of temperatures since about 2000 (see Figure 5).

This paper appeared at exactly the right moment for me. I was looking for the right punch line for a proposal which revolves around water vapor and satellite observations, and here it was. It arrived with a strong climate effect for the stratospheric water vapor and could lead to especially important open questions (always important for proposals: open questions!). Does the observed variability of stratospheric water vapor have something to do with greenhouse gases and global warming, or is it simply a natural decadal variability? At least some anthropogenic influence on stratospheric water vapor could hardly be disputed, namely, that if the amount of methane in the stratosphere increases so must its oxidation product water. But besides that? Solomon leaves this issue truly open, plenty of fodder for future research.

And the (pseudo) skeptics? And the press? Although Solomon's paper was published in Science I had not thought that the issue would interest any of the Nutters, let alone the daily press. But something is afoot. Here are some funny interpretations from Lala-land:
water vapor drives the climate, not CO2, a new study authored by Susan Solomon, could explain why atmospheric carbon is not contributing to warming significantly.
Certainly I was the most surprised to read about Solomon's work in an article in Die Zeit by Jurgen Krönig. And it was not in an article that dealt specifically with stratospheric research, but in one of many articles, in which each and every journalist who had never peeked into the IPCC report, expresses their dismay that on page 800 or whatever, the disappearance of Himalayan glaciers has been incorrectly predicted. In the midst of this discourse, which appears to discuss the abolition of IPCCs, and standing Pachauri up against the wall on the spot, to my complete surprise the following sentences appeared:
New findings could worsen the situation further. A new study published in Science, shows that role of water vapor, the main greenhouse gas, has been neglected in the IPCC’s climate models the IPCC. In the stratosphere, the concentration of water vapor decreased by 10 percent which reduced the temperature rise by 25 percent "
Did Solomon say that the climate models neglected water vapor? Sure and Roger Federer neglects his forehand. And as if by chance, Krönig refers to the period from 2000 to 2007 (with the decline of stratospheric water vapor and the corresponding forcing -0.1W/m2, see above) without further consider that, according to the same paper from Solomon the overall effect of stratospheric water vapor from 1980 until today (see Figure 5 above) is that greenhouse gas driven warming has intensified. In this way Krönig contributes mightily to the advertising campaign for January's climate shock of the month competition.



Anonymous said...

I look forward to reading this in more detail Eli, glad to see that you have tropospheric folding in there!


carrot eater said...

I was going to snark that this commentary was longer than the paper itself, but then I found it was worth reading. But what I was really hoping for was some thoughts on the open questions as to the nature of this variability.

You've been promoted from old hare to Godfather, I see.

David B. Benson said...

Thnaks, I found this quite helpful.

But there remains the reminder for paleoclimatology that the Eemian interglacial was about 2--3 K warmer, so any imputed negative feedback has to be small indeed.

DeWitt said...

Given that the temperature in the stratosphere increases with altitude, doesn't the physics of water vapor/liquid equilibrium not merely allow, but require the water vapor concentration to increase with altitude? It certainly increases with altitude over the ice caps in winter when there's a temperature inversion.

carrot eater said...

DeWitt: Why would it be required? There's no ocean up there that's in local equilibrium with the air. There isn't 100% relative humidity, or anything close to it in the stratosphere. Thermodynamically, the RH can be whatever it wants, upto 100%.

Anonymous said...

Little Mouse finds the tropopause fascinating. It is not that the temperature changes abruptly, but that the rate of temperature change, changes abruptly.

Although in the graphs of actual temperatures Little Mouse has seen there are two discontinuities. Like an aircraft with the tip of the nose missing.

Could the tropopause be a layer in its own right? Or was that a once off and the Mouse has seen the same graph again and again.

Curious how nothing just a blip in the rate of temperature change provides such an effective barrier. that needs a big event to pass through. (Best not to do that at exactly the speed of sound.)

Chip Knappenberger said...


Hoffmann writes “Did Solomon say that the climate models neglected water vapor? Sure and Roger Federer neglects his forehand.”

Well as to water vapor in the stratosphere, Solomon et al wrote: “current global models are limited in their representations of key processes that control the distribution and variability of water within the stratosphere” and “[c]urrent global climate models simulate lower stratospheric temperature trends poorly.” I think Federer has more than a “limited” handle on his not-so-“poor” forehand.

Anyway, Solomon et al. showed that stratospheric water vapor variability in the stratosphere is currently not well understood (or modeled) and may have a fairly large impact on multi-decadal surface warming trends. As such the IPCC et al. should perhaps be a bit more careful is assigning their likelihood descriptor to recent warming trends. In other words, that significant influences on the surface temperature trends are still being uncovered should temper the IPCC et al.’s enthusiasm for the certainty of the attribution of particular values. See here for how they can quickly get in trouble.

-Chip Knappenberger

Anonymous said...


Regarding the models, you are splitting hairs. I think Eli's point is that they do model the process in question, albeit imperfectly. They most certainly do not ignore those processes. Just b/c they are imperfect that does not nullify their value.

You use words like "significant" and "fairly large" to describe the relative impact of SWV on the net forcing. A simple scale analysis shows otherwise. In 2008 the estimated radiative forcing of GHGs was 2.74 W m-2, see:

In contrast, the forcing from SWV was estimated by Solomon et al's model to be at most 0.25 Wm-2. There is an order of magnitude difference. Additionally, about 6-7 years of for the forcing from GHGs to exceed that forcing modeled by Solomon et al.. It is thus not surprising that the long term trend in Fig. 3 is positive.

IMHO, this finding by Solomon et al. has no bearing on climate sensitivity estimates to doubling CO of ~+3C. Those sensitivity estimates are based primarily on paleo climate data which quantify the net forcing for the entire system/biosphere. That is, they implicitly take into account factors such as forcing from SWV.

Can we ignore the forcing from SWV? No. Is it a major or significant player? No.

The real interest in Solomon's work is what caused the drop off in SWV circa 2001. Is this natural or somehow related to anthro activities? My suspecion is that it is decadal variability, the record is unfortunately too short to make any definitive statements.


Anonymous said...

MarkeyMouse says: "This paper appeared at exactly the right moment for me. I was looking for the right punch line for a proposal which revolves around water vapor and satellite observations, and here it was. It arrived with a strong climate effect for the stratospheric water vapor..." Hello in La La land, why do you think she concocted it?

"Does the observed variability of stratospheric water vapor have something to do with greenhouse gases and global warming, or is it simply a natural decadal variability?" Water vapour always increases with warming, it is part of the self regulating temperature mechanism.

Anonymous said...

"Is it a major or significant player? No."

That seems rather categorical and unqualified. On what do you base it?


Anonymous said...

NewReader, based on the quantitative scale analysis in my post @10:27 am. Had you read my post properly, you would have seen that it is a qualified statement.

Eli and others in the know concur.


carrot eater said...


Be careful.

There are bookkeeping issues here, and one must keep them straight.

Your 2.7 W/m^2 RF for greenhouse gases is for 2008 relative to 1750.

Solomon's 0.24 W/m^2 for stratospheric water is for 2000, relative to 1980. I think. If I'm wrong on that, then this entire comment is wrong.

If you compute total greenhouse gases in 2000 relative to 1980, you get 0.726 W/m^2.

Put that way, the stratospheric water vapor looks like it can be fairly significant on these times scales.

But we're again left with a matter of bookkeeping - what is this variation in stratospheric water? Is it forcing, feedback or a result of internal variability? This question is critical to understanding how important this work is.

To the extent that stratospheric water comes from methane oxidation, stratospheric water is already included in the models as a forcing. In GISS E, that's good for 0.06 W/m^2 in 2000, relative to 1750. But this paper is about water transported up from the troposphere. I haven't had time to see how much that varies in the models.

But now, I do appreciate that miniscule changes in strat water are worth looking at. Eyeballing from the graph, it went from 3.5 to 4.something back to 4 ppmv. That doesn't sound very important, but one must always do the sums to check.

carrot eater said...

Following up, you can see the relative significance more clearly in Fig 3A and then Fig 3B of the paper. It's still minor, I'd say.

Well, more clearly if I could figure out what the hell it was showing; I'm a bit confused by how the red line is calculated, or exactly what it represents.

The impact on the temperature trend doesn't really seem that important, if you look at the figure. On these time scales, you might not observe the difference, anyway; there is too much noise.

If the stratospheric water content just fluctuates around 4 ppm like that, it really would not matter at all in the long run. Given the lags in the climate system, it'd just wash itself out.

Anonymous said...

Thanks for pointing that out carrot-- good points. I did not read the text that went with the table correctly. Regardless, 0.25 is still about 33% of the instantaneous (?) GHG forcing.

The point is that while the variation of SWV may be of note on short time scales, just as internal climate modes can be, in the long term it is not of significance. And, ultimately, that is what we are concerned with when it comes to AGW, the long term.

carrot eater said...

That's probably fair, so long as it keeps bouncing around 4 ppm. I wish we had a longer observational record of it.

I do think some of the statements in the paper are not well justified by the results, and I don't like how Fig 3B (Fig 5 above) is constructed. I wonder if their dotted line and blue line would both fit within the envelope of possible outcomes in a coupled GCM.

Anonymous said...

Carrot, others have also pointed out that some of the statements made in the paper have not been written terribly clearly, thereby making them open to misinterpretation, especially by the denialati.


Chip Knappenberger said...


I totally agree that the paper is written very nebulously and loosely . I have a great deal of difficulty with Figure 1c. The SST axis is flipped (although it doesn’t mention this anywhere I can find), and so are the SST/water vapor correlations negative or positive? It is described in the text as:

"As discussed above, recent observations have suggested a correlation of the post-2000 stratospheric water vapor decrease to sea surface temperature changes near the tropical warm pool region and associated cooling of the cold point that governs water vapor input to the stratosphere in the tropics (Fig. 1C). However, the relationship between SSTs in the warm pool region and stratospheric water vapor changes character (from negative to positive short-term correlations) from 1980–2009, suggesting that other processes may also be important, or that the correlation may be a transient feature linked to the specific pattern of SSTs at a given time rather than to the average warming of SSTs around the globe."

It changes character from when to when (Figure 1C doesn’t depict water vapor back to 1980). What is “short-term”? 2 years, 8 years, 15 years? Are the correlations different at different timescales? Is there any analysis to back this up? This is all OK for *Science* magazine? I have no problems with a paper published that describes a potentially important effect but that can’t fully describe why it is happening, but in this case, it is *very* difficult to even understand the description of why the authors don’t understand!


carrot eater said...

I also thought that discussion was a mess. If a correlation changes sign over time, is it really a "correlation"?

Dano said...

Splitting hares as a dissembling tactic notwithstanding, let's just say back in the day if we were forecasting that shift and saw that vortmax coming our way, we'd be wagering on precip totals and wind speed advisory maxes.

I did a shift in Europe where we had something like that coming (way before Internets and our data weren't that good, alas), and we had aircraft tip over on our airfield. I think it was before Chernobyl so it didn't do any good at clearing out the cancer from our area...



Martin said...

Hi Eli,

Excellent translation.
Just one comment: You translated Klimaschmock as climate shock.
I'm not sure whether the word schmock is a proper English word. It could be translated as a hack.

Georg anoints one every month and at the end of the year there is a poll to elect the climate schmock of the year.

PS Now you're a godfather, can you make Senator Inhofe an offer he cannot refuse?

john S. A. said...

The interesting thing is, in my opinion, that people that tend to suggest that climate is not sensitive enough to GHG forcing (CO2 forcing is negligible, in their wording) cling to a -0.2 W/m2 forcing now to discredit IPCC's work. I can't understand why they say this forcing is important (and I think it is) but the forcing associated to GHGs is not (and I think it is too).

So ... the climate has different sensitivities to radiative forcings depending whether the infrared photon arriving to surface was emitted by WV or by CO2?

Curious position

Cheap Lots in Samara Costa Rica said...

These articles are fantastic; the information you show us is interesting for everybody and is really good written. It’s just great!! Do you want to know something more? Read it...: Great investment opportunity in Costa Rica

antique hand tools said...

Everyone is aware that it cools as you rise higher in the atmosphere and as a consequence it has to become dryer.

breast engorgement said...

It is generally assumed that most of the water vapor found in the stratosphere passes through the tropical tropopause which is the coldest point of the troposphere, into the stratosphere.