A Fresh Look at Clouds, and Heat, in the Greenhouse

Here’s a deeper look at some of the points explored by Justin Gillis in his article earlier this week on the persistent questions, and hints of answers, surrounding how clouds will respond as concentrations of greenhouse gases continue to rise. (The graphic that accompanied the article is well worth a deeper look, too. Click here to expand on what you see to the right.)

Much of the heated discussion of the article on the Web focuses on its treatment of the skeptical stance of Richard Lindzen, the Massachusetts Institute of Technology climate scientist who has long been a hero of those fighting efforts to restrict greenhouse gases.

The Lindzen focus is a distraction, to my mind. Anyone thinking that the erosion of Lindzen’s credibility will somehow build societal enthusiasm for cutting greenhouse gas emissions is probably overly optimistic. This was expressed bluntly in Mother Jones by Kevin Drum, who wrote that clouds (and Lindzen) aren’t “deniers’ last hope,” adding, “Human greed and self interest are their last hope, and there’s very little chance of that diminishing anytime in the near future.”

More importantly, though, it’s a distraction from the issues surrounding how clouds might amplify or blunt human-driven warming. The range of possible consequences of this warming runs from manageable to catastrophic. The uncertainty has been durable, resisting many years of concerted research and analysis, but Gillis wrote that there are signs of progress. (Gillis dives in deeper on the science in a subsequent Green Blog post.)

After reading the initial article, I sent a query to Anthony D. Del Genio, a NASA climate modeler long focused on the cloud question. In Gillis’s story, Del Genio was the prime voice of optimism. Here’s our exchange (with some e-mail shorthand cleaned up):

Q.
It always amazes me how even a very long climate article (by me, Justin Gillis or anyone else) still only seems to scratch the surface. I was wondering what findings or research prospects led to your tone of optimism [on clarifying the role of clouds in greenhouse-driven climate change]?

A.
Justin asked me the same question. Here was my answer, copied from an email I sent him after we spoke:

I feel that we’re * on our way * (not there yet – but starting to see a path forward and a few signs of success) to doing a lot better because:

- We have realized how sensitive convective storms (thunderstorms and other heavy rain events) are to the humidity of the air around them, and we are beginning to account for that in our climate models

- These storms should penetrate higher as climate warms according to the models, a positive feedback, and satellite data looking at cloud height changes over El Nino time scales show something similar and show the models getting that about right also, for physical reasons we think we understand

- The midlatitude storm tracks are predicted to shift poleward as climate warms (an expansion of the tropics), a positive feedback, observations are beginning to show evidence of that too, and there are several good explanations for why that should occur

- Now that we have almost 3 decades of satellite cloud data, people (e.g. Joel Norris of UC San Diego) have begun to see an overall trend of more cloud near the equator, less cloud in the subtropics and midlatitudes, and the models are beginning to reproduce that general pattern

A big reason we’re not there yet is the subtropical cloud decrease part – satellite trends seem to show it, many models (including ours) predict it, but I think the physics of low subtropical marine stratocumulus and shallow cumulus clouds is not yet represented well enough by the models to be confident that if we are getting it right, we are getting it right for the right reason.

But I’m optimistic about the rest of it because we now have viable physical mechanisms for why the other types of clouds should change in a particular way as climate warms and some tentative observational evidence that this is correct.

That was my attempt to summarize the state of things in bullet form (as close to bullet form as I can get, I tend to be wordy), but if you’d like to know anything about any of these points, let me know.

Q.
This is incredibly helpful additional information…. I still see a question here, though. You say we’re “on our way,” but where? The description of tropical shifts you include in this note doesn’t get at the question of the resulting global (or regional) mix of positive and negative influences on the planet’s energy balance from such shifts in cloud patterns.

Are there already hints in this mix of findings that the climate’s sensitivity to a CO2 buildup is at the higher, lower or mid-range of what’s possible? Or are you saying there are hints of research directions that could, in the long haul, narrow the bounds of the question a bit?

A.
Right. I say that we’re on our way, but not there yet, because a decade or so ago, (a) the only observational evidence was about other types of current climate variability in clouds rather than trends, (b) no one had yet demonstrated a relationship between that behavior and climate change, (c) there were few if any accepted physical mechanisms for why specific types of regional cloud feedbacks occur. Now we have a couple of mechanisms that seem realistic for several specific types of changes such as the cloud height feedback, and we have some observational confirmation of a general latitudinal pattern of feedbacks that many models seem to get, more so than was the case a decade ago. So we’re beginning to feel we are doing *something* right, and we are starting to zero in on the most important questions that still need to be addressed in a way that I don’t think was true a decade ago.

But as you say the spread is still there because the relative magnitudes of the different regional feedbacks differ among the different models, so one model’s mix of + and – feedbacks will differ from another. So the fact that a pattern seems to be emerging, in both observations and models, which was not the case a decade or so ago, is encouraging. And the physics behind some aspects of those patterns is beginning to be understood. But not all the mechanisms are well enough understood yet (e.g., low clouds as probably the biggest nagging issue), so the spread is still there.

And of course just like cars, not every climate model is good, yet to date we just quote the entire spread among all models. The challenge is to define meaningful metrics to evaluate the models (probably different ones for different climate questions — climate sensitivity and southwest U.S. drought may respond primarily to very different things). This much harder to do for climate than for weather.

The community hasn’t yet done enough of that hard work to accomplish that. Instead, the tendency has been to just look at static mean geographic patterns of things, which has been shown over and over again not to have much predictive value, with a couple of exceptions (e.g. see a nice 2010 paper by Trenberth and Fasullo [summary here] that relates one chronic model mean cloud error to climate sensitivity). You have to watch clouds change in response to a change in the environment – a dynamic view – to get better insights. What the emerging patterns and mechanisms do for us is to give us a research path to perhaps define truly useful metrics of that sort, not easily calculated metrics that do not help us. Whether that’s long haul or short haul, hard to tell…not only for science reasons.

As a scientist, I of course wish some of what we know about clouds now could get into print, but for most people it seems to be an arcane subject. But it’s not the hunt for the Higgs boson, which everyone now seems to know about.

- New York Times