In Climate Science, Beware the Conversion Narrative

Okay, so: Yesterday I gave a good, long, hard read to this paper, just published in the Bulletin of the American Meteorological Society by MIT hurricane guru Kerry Emanuel and some colleagues. It was incumbent upon me to do so for two reasons:

1. Emanuel is one of the main personages in my book Storm World, which chronicles much of the hoopla that occurred after he and another team of scientists published twin 2005 papers suggesting that global warming had markedly increased the intensity of the average hurricane;
2. this latest paper has inspired press reports suggesting that Emanuel is now changing his tune, backing away from his previous conclusions.

I read the paper – describing a new technique for studying hurricanes under climate change via complex computer models – and I have to say, I'm a bit mystified about precisely where the supposed conversion or backing away lies.

In essence, Emanuel and his colleagues generated a large number of synthetic storms – 2,000 for each hurricane basin of the world – and then compared their fates by using a suite of global climate models to simulate first the climate of the late 20th century, and then of the late 22nd. And the results were all over the place – different basins of the world saw very different changes in storm activity depending upon the model used. But there were some general trends, to wit:

...there are substantial increases in power dissipation in the western North Pacific and decreases in the Indian Ocean and through the Southern Hemisphere, while the tendency is somewhat indeterminate in the eastern North Pacific, with large variability from one model to the next. Four out of the seven models show appreciable increases in the North Atlantic, with the remaining three showing no change or decreases ... There is a general tendency toward decreasing frequency of events in the Southern Hemisphere, but the percentage change in frequency is not usually as large as the decrease in power dissipation ... Changes in overall frequency in the eastern North Pacific are indeterminate, but six out of seven models show increases in the western North Pacific and five out of seven models show increasing frequency in the North Atlantic. Overall, the tendency of storm frequency is somewhat indeterminate in the Northern Hemisphere, but declines in the Southern Hemisphere.

That's quite a lot to process, so let's break it down even more: Using this particular technique, it appears that global warming should decrease the total number of storms while strengthening them in some locations (notably, the Atlantic and the Northwest Pacific). Incidentally, this result consistent with other computer modeling studies. But as the new Emanuel paper notes, there are many reasons to question whether those studies are yet at a point where they can be fully relied upon:

...large model-to-model variability and natural multidecadal variability within at least some of the models also suggests large uncertainty in such projections, reflecting the uncertainties of climate model projections in general and the influence of natural variability.

And now, let me give you my take on what is really going on here. In 2005, Emanuel published a paper suggesting a dramatic increase in global hurricane intensity that had already happened. This triggered a huge debate and ever since then scientists – Emanuel and a number of others – have been doing follow-up work to try to determine the reliability of that contested result and just generally to advance knowledge on the hurricane-climate front. The follow-on studies have used a variety of methodologies, with computer modeling of the type described above being only one approach. And generally, when you look at all the work taken together, you draw the following conclusions:

1. Hurricanes are going to change due to global warming, and probably already have.

2. We don't know precisely how they are going to change, because there is going to be variation across different parts of the world and there could well be offsetting changes between frequency and intensity.

3. We have a general expectation of increased storm intensity (which goes back to work by Emanuel decades earlier).

4. We have an increasing expectation, based on climate modeling studies, of decreased total storm frequency (but note that this may not be true in all parts of the world).

To be sure, this leaves an unresolved issue: What is currently up with the Atlantic? After all, there has been a dramatic increase in storm frequency and intensity there over the past few decades. Is this global warming at work?

Emanuel's 2005 paper said that it was. By contrast, if you read it carefully, this latest paper seems open to the possibility that this 2005 result was wrong. After all, the dramatic Atlantic ramp-up that we've seen over the past 25 years or so hardly continues forward in a linear way in the climate modeling studies that Emanuel has more recently performed to examine our global warming future. And if the models don’t show a continuing dramatic response to even more global warming than we've seen so far, then how could what we've seen so far be caused by global warming?

That's a fair question to ask, but there's a catch: Though he is now doing studies using global climate models, I don't think Emanuel completely trusts them. As the new paper explains: "...[either] the greater part of the large global increase in power dissipation over the past 27 yr cannot be ascribed to global warming, or ... there is some systematic deficiency in our technique or in global models that leads to the underprediction of the response of tropical cyclones to global warming." And Emanuel himself is a leading proponent of the idea that all of these global models are systematically deficient because they fail to take into account how hurricanes themselves affect the climate system through their mixing of the oceans and contribution to the transport of heat through them toward the poles.

The upshot? Kerry Emanuel is an open-minded guy who is working through a very complex problem. A 2005 study opened his eyes, shocked him, and since then he has been trying a variety of techniques to refine his understanding. He admits that 2005 study may have been wrong – any good scientist would – and he continues to study the problem, trying to explain the dramatic data we've seen from the Atlantic and reconcile it with what theory and models would predict. This latest study adds to the total well of knowledge, but doesn't, so far as I can tell, really represent any major position change for Emanuel.

Note: I am emailing this blog post to Kerry Emanuel to see if I've gotten anything wrong or misinterpreted. Also, the author of the article suggesting that Emanuel had changed his views, Eric Berger of the Houston Chronicle, is one of our very best hurricane-climate reporters, and I take his work very seriously. I'll also be sending this piece to him. A subsequent post, if necessary, will work through any responses they may have.