Saturday, August 19, 2006

More on the water vapor feedback

The water vapor feedback is one of the most important processes in our climate. In fact, the feedback is responsible for a significant part of the warming predicted to occur over the next century.

A reader asked an interesting question in response to a recent post on the water vapor feedback: What peer-reviewed evidence exists for a positive water vapor feedback? What about a negative water vapor feedback?

Let's take the negative feedback first. I did a quick search on the Web of Science and found out these statistics:
70 papers contained the phrase “water vapor feedback”
18 of them contained “negative” and “water vapor feedback”

If you go through the abstracts, you find that only four articles talk about a negative water vapor feedback (in other abstracts, the word “negative” was modifying another phrase). I’m adding a fifth paper that was not flagged in my search because it was published in 1990, before the WOS included abstracts. Also, I’m dropping one paper for reasons I won’t go into here.

Here is the resulting list of peer-reviewed literature on the negative water vapor feedback:
1. Lindzen, R. S. (1990), Some coolness concerning global warming, Bull. Amer. Meteor. Soc., 71, 288-299.

2. Sun, D.-Z., and R. S. Lindzen (1993), Distribution of tropical tropospheric water vapor, Journal of Atmospheric Sciences, 50, 1643-1660.

3. Sun, D. Z., and R. S. Lindzen (1993), Water-vapor feedback and the ice-age snowline record, Annales Geophysicae, 11, 204-215.

4. Lindzen, R. S., M.-D. Chou, and A. Y. Hou (2001), Does the Earth have an adaptive iris?, Bull. Am. Met. Soc., 82, 417-432.

Hmmm. There’s a pattern here that I just can’t quite figure out. Just joking. The pattern, of course, is that only Dick Lindzen has been able to publish a paper arguing for a negative water vapor feedback. But if you look carefully at the papers, they make a much weaker argument than you might expect.

Paper 1 is considered wrong by everyone, including Lindzen. In fact, paper 2 was written to correct the defect in paper 1. Paper 2 mainly discusses how tropospheric water vapor is regulated. Its discussion of the water vapor feedback is limited to a short discussion of how a negative water vapor feedback might plausibly arise, but no evidence of such a feedback is given. I cannot remember what’s in paper 3, and I can’t find a copy. However, based on the date and author list, it’s likely that it contains much of the same as in the other 1993 Sun and Lindzen paper.

In paper 4, Lindzen resurrects his idea from paper 2, and provides some data to argue that indeed a negative water vapor feedback does exist. Unfortunately (for him), subsequent tests of other scientists failed to verify this idea. At the present time, there is virtually no support in the community for it. Lindzen still gives talks on this and claims that the iris hypothesis is still viable. However, as far as I can tell, no one, including Lindzen, is working on it, so that indicates exactly how vibrant the idea is.

So let’s summarize. There have been a very small number of articles written that argue for a negative water vapor feedback. Virtually all have been written by Dick Lindzen. None have stood the test of time.

Now let’s turn to the other side. What evidence is there for a positive water vapor feedback. First, there are 25 papers that contain “positive” and “water vapor feedback”. Many of these papers are recent (written in the last 2-3 years). Going through the abstracts reveals that most of these papers argue in favor of a water vapor feedback. Many conclude this directly from data, not from any type of GCM analysis. In addition, the papers were written by a large number of different scientists.

Here are a few positive-feedback papers that you might want to take a look at (just a small subset of the literature):
1. Minschwaner, K., and A. E. Dessler (2004), Water vapor feedback in the tropical upper troposphere: Model results and observations, J. Climate, 17, 1272-1282.

2. Minschwaner, K., A. E. Dessler, and P. Sawaengphokhai (2006), Multi-model analysis of the water vapor feedback in the tropical upper troposphere, J. Climate, accepted.

3. Soden, B. J., et al. (2005), The radiative signature of upper tropospheric moistening, Science, 310, 841-844.

4. Sherwood, S. C., and C. L. Meyer (2006), The general circulation and robust relative humidity, J. Climate, in press.

5. Dessler, A.E., and K. Minschwaner, An analysis of the regulation of tropical tropospheric water vapor, J. Geophys. Res., submitted.

[pre-prints of the Sherwood paper can be found on his web page; if you want a pre-print of paper 2 or 5, let me know].

These are just a few of the more recent papers. A bunch more exist, written by different scientists using different data.

Thus, the evidence in favor of a positive water vapor feedback is strong, with multiple peer-reviewed analyses reaching this conclusion. The evidence that the feedback is negative is weak: only Dick Lindzen argues it, and his arguments have been roundly rejected by the scientific community.

Also note that Bill Gray hasn’t published anything on this topic. That’s because he does not have a testable hypothesis nor any data, both of which would be required. Nor can he criticize the published literature on this subject because he has not read these papers.

The upshot: we can conclude that the scientific community agrees that the water vapor feedback is positive. Arguments to the contrary are distortions of the science.

[Note: I recognize that these WOS-type analyses have their pitfalls. But as a scientist who publishes in this area and who has read (I think) all of the relevant literature, I can attest to the fact that this WOS analysis has got it right: there is just about zero evidence to support a negative water vapor feedback.]


Anonymous said...

Given an area of atmosphere (i.e. from surface to space), and a given amount of water vapor, there will be a certain % of cloud cover (type and altitude to be determined). If the surface (with 70# water) heats delta T this will lead to additional evaporation of water vapor delta H. Now, the question is, do you agree or disagree that this will lead to an increase in cloud cover delta C or not? This would seem to be the first question to be discussed.

Dave Dardinger

Anonymous said...

I suppose that the more interesting questions to investigate should be "positive" and "water vapor feedback" vs. "negative" and "cloud feedback". But the latter is troubled by some ambiguous titles like "clouds: a negative or positive feedback"...

But more serious: I suppose that most scientists agree that water vapor gives a positive feedback, and that most scientists agree that more (low level) clouds give a negative feedback. Where there is much disagreement is if (A)GW induces more or less clouds. That is the main cause of the rather wide range of model results for 2xCO2.

But many, if not all, models don't capture cloud cover in the tropics (and the Arctic) that well. Especially the reduction of cloud cover in the period 1985-2000 in the (sub)tropics, caused by an increase of the Walker/Hadley cell circulations, a drying up of the upper troposphere and a change in radiation balance: 2-3 W/m2 more insolation, but 5 W/m2 more radiation to space. Or an overall loss of ~2 W/m2 TOA. Or a negative feedback to higher sea surface temperatures (but what is cause and effect here?) of the same order as the sum of all extra GHGs since the start of the industrial revolution for (more than) halve of the earth's surface...

See: Wielicki ea. and Chen ea.
Recently confirmed by: Pinker ea.

That models don't capture the observed change in cloud cover/radiation even not for the 60N-60S area can also be found at: Allan and Slingo.

Andrew Dessler said...


I'm gratified to see that the blogosphere is coming around to the idea that the WV feedback is positive.

That said, I agree that the next big uncertainty is the cloud feedback. In fact, that's what I'm working on now. If you read Soden and Held's new J. Clim. paper, you see that most models include a slightly positive cloud feedback. Experimentally testing that is going to be a challenge, but challenges like that are what data analysts like me thrive on.


Anonymous said...


Depends where you look. In the subtropics, the faster Hadley/Walker cell circulation increases cloudiness in the ascending air column, but decreases cloudiness in the descending column, the net result is a slight reduction in the tropics (Walker cell) and more pronounced reduction in the subtropics (Hadley cell).

This is mainly for high-level (cirrus) clouds, which in general have a warming effect (little effect on SW reflection, larger effect on LW reflection back to earth).

The reduction of cirrus clouds induces more insolation and higher sea surface temperatures, mainly in the subtropics. But the net radiation balance at the TOA is negative, because most of the increase in IR radiation back to space is at higher altitudes. Thus somewhere in the total balance, more heating of the (sub)tropics leads to more loss to space for the globe, which is far larger than the initial heating: the increase in GHGs over the period of interest (1985-2000) theoretically gives ten times less extra forcing (including WV feedback) than the observed extra heat loss in the balance...

Lindzen did think about the same mechanism, but on a much smaller scale: within the warm pool, ascending air/storm clouds vs. descending dryer air. That is much harder to (dis)prove, as there still is discussion where the border is between (thin cirrus) clouds and clear skies...

The Norris findings confirm and extend what Chen & Wielicki ea. have found by adding surface observations of cloud cover over previous periods and higher latitudes.

Anonymous said...


No matter what heats the ocean surface, this leads to faster air circulation, less cirrus and more loss of heat to space, based on the result of the observations as discussed in my previous mail.

Even internal oscillations, like an El Nino (1998) leads to ~7 W/m2 more net loss of energy to space, due to higher sea surface temperatures. Thus IMHO, if models include a small positive feedback by clouds, they are not reflecting reality.



Anonymous said...

In addition, J. Norris has a paper in preparation about cloud trends vs. climate change.

On page 58, there is a calculation of cloud feedback, assuming that the change in cloud cover is solely a response to increased forcing. The net response is -0.8, which is a very strong negative feedback...

Of course this is the maximum response, if nothing else is influencing cloud properties/cover, but important enough for further investigation.

The paper in preparation gives a very detailed overview of cloud properties and influence over time.