Major New Paper on Hurricanes and Global Warming

In order to get a tropical cyclone spinning, a lot of things have to go right (or wrong, depending upon your perspective). First, you need a location that's warm but also a certain distance north or south of the equator. In places too close to latitude zero, winds won't swirl inwards towards an area of low pressure to create a cyclonic rotation (a phenomenon known as the Coriolis effect). Second, you need a temperature gradient between the warm ocean surface and the cooler atmosphere above it, a situation that's favorable to what meteorologists call convection (the transfer of heat upward with rising air).

And that's not all. You also need a region with low levels of what's known as vertical wind shear. When winds blow in different directions at different altitudes, they can tear apart a hurricane's structure. Furthermore, you need high sea surface temperatures--above 80 degrees Fahrenheit. That's particularly crucial, since the heat energy stored in the ocean fuels the hurricane, first entering the air through evaporation and later being released higher in the atmosphere when water vapor condenses into rain and clouds. This upward transfer of energy explains how a hurricane's heat pump--or heat engine, depending upon the metaphor you prefer--really gets chugging.

And even with all of these conditions in place, hurricanes still won't form without some initializing disturbance or "trigger," such as an African easterly wave. No, these aren't the kind of waves that that surfers catch. It's better to think of them as wave-like patterns of air, often accompanied by thunderstorm clusters, that ride easterly winds across the Atlantic and often stir up hurricanes during their passage.

All of these factors--and others as well--help determine whether or not a hurricane comes to exist in the first place. However, when scientists suggest that global warming might be intensifying hurricanes, it's generally changes in sea surface temperature (SST) that interest them most. The basic argument runs like this: All other factors being equal, higher SSTs will strengthen storms and lengthen their duration by providing a greater energy source for them to draw upon. This argument is rooted in our modern thermodynamic understanding of hurricanes, but it also makes a kind of intuitive sense. After all, we know that storms weaken over cold water, or over land, because they lose their energy source. No wonder running over warm water, and especially deep warm water, strengthens them.

So when two scientific papers came out last year linking global warming to an increased number of hurricanes reaching Category 4 and 5 intensity levels (1, 2), it's no surprise that the suspected mechanism driving this outcome was a global increase in sea surface temperature (an increase linked to global warming). However, neither of the papers did an in-depth analysis to demonstrate that SST change was, in fact, causing the observed change. They merely noted correlations between hurricane strength and increasing SST.

But now, in the latest issue of Science, several members of the Webster group at Georgia Tech--which produced one of those initial papers (2)--have taken the next logical step (3). Their new paper seeks to determine whether higher SSTs are, indeed, causing a growth in hurricane strength--or, alternatively, whether that change can be attributed to changes in other important factors like vertical wind shear, atmospheric heat and humidity, or something called "zonal stretching deformation" (which I don't completely understand, but which apparently relates to the Coriolis effect again). And sure enough, they found that all the evidence pointed towards SSTs.

The Georgia Tech scientists report that of the four factors they studied, only SST had been trending upward consistently in all six of the ocean basins considered. For the other three factors, no consistent global trends were observed (though there were trends in some of the individual basins). And so, after some complicated statistical/"information theory" jazz designed to figure out the relative importance of these variables--there should be an award to the journalist or blog denizen who can explain this part of the study in layman's terms--the paper (of which I have a copy, not sure it's available online yet) concludes:

The implication of these results is that the strong increasing trend in [the number of Category 4 and 5 hurricanes] for the period 1970-2004 is directly linked to the trend in tropical SST and that other aspects of the tropical environment, while influencing shorter term variations in hurricane intensity, do not contribute significantly to the global trend of increasing hurricane intensity.

So what's the take-home point here? Well, this paper would certainly seem to strengthen the conclusions of previous studies about a link between hurricane strength and global warming. More specifically, the new study, by attempting to rule out other variables that affect hurricanes, adds support to the contention that the link between global warming and hurricane intensity hinges upon rising sea surface temperature. Hurricane theory had long pinpointed SST as the chief source of storm energy, but now a statistical test seems to back that up.

To be sure, we'll have to see whether these results hold up after criticism. My guess is that they will gain ample media coverage, and that this coverage will contain critical reactions from scientists who remain skeptical of the initial studies linking hurricanes to global warming. But of course, that's how science is supposed to work. Indeed, I'm told that the original idea for the new study actually arose because the authors were responding to criticism of their previous study published in Science (2). That exchange, I understand, is currently in press.

In any case, I'm no scientist, but allow me to just wade in here with one slight comment of my own (I hope I can explain this right). As I've noted above, many factors control whether hurricanes develop and how strong they become, and it's very hard for scientists to say precisely how some of those could change in the future as a result of global warming. Take one of the parameters examined in the latest study--atmospheric temperatures, which, if rising, could suppress hurricanes. The latest study suggests that this isn't a major factor yet, but it certainly could become so. After all, the lower the temperature difference between the ocean and the upper troposphere, the harder it is to get the upward heat transfer going through convection. And there are good reasons to think that global warming can or will trigger upper tropospheric warming. So if that happens, it might offset some of the importance of the increase in SST on hurricane strength--although I'm not sure anyone thinks it would totally offset the significance of the SST increase. But it does go to show you how complex this subject is.

In any case, that's my two cents based upon some reporting that I've been doing. From a political perspective, the upshot is that we can expect the rather testy hurricane-global warming debate to flare up again over this paper, so that will be interesting to watch. Meanwhile, if anyone can explain the new paper's stuff about information theory in English and without equations, I strongly urge you to do so in the comments. Moreover, if I've explained any of the above science wrong, which is likely, please correct me--that's what blogging is for....

1. Emanuel, "Increasing destructiveness of tropical cyclones over the past 30 years," Nature, Vol 436, August 4, 2005. (PDF)
2. Webster et al, "Changes in Tropical Cyclone Number, Duration, and Intensity in a Warming Environment," Science, September 16, 2005, Vol 309. (link)
3. Hoyos et al, "Deconvolution of the factors contributing to the increase in global hurricane intensity," Science, March 17, 2006, Vol. 312. (abstract)

More like this

I’m sure the measurements are still being checked and adjusted but it is clear that Typhoon Haiyan/Yolanda was one of the most powerful tropical cyclones (termed “Typhoon” in the western Pacific) ever recorded. There are several ways to measure how big and bad a tropical cyclone is including it’s…
Tomorrow I am off to New York for this meeting on hurricanes and climate, sponsored by Columbia University's International Research Institute for Climate and Society. I'm very appreciative that Columbia-IRI has allowed me to attend the event, which will feature presentations from Kerry Emanuel,…
Storm World: Hurricanes, Politics, and the Battle Over Global Warming by Chris Mooney Harcourt: 2007, 400 pages.Buy now! (Amazon) At 2:09 am on September 13, 2007, Hurricane Humberto made landfall just east of Galveston, Texas--still the site of the deadliest natural disaster in US history, the…
It may seem a strange question to be asking in a season that, so far, hasn't yet seen an Atlantic hurricane. But while the weather in any given year can be tricky and unpredictable, there's no doubt that we're currently in an active period for Atlantic storms in general--and now, a new paper (PDF…

There should be some interesting modeling going on about this. Even if tropospheric temperatures increase, thus perhaps suppressing formation of hurricanes, the energy contained in the warm water has to go some place. There must be a transfer of that heat to cooler regions. What form will it take?

By Mark Paris (not verified) on 16 Mar 2006 #permalink

Good question. Nobody knows, of course, but nobody seems to think that hurricanes would just die out completely. It might be that it would be harder for them to "break through" the warmer, static atmosphere, but that once they did so, they would be incredibly powerful due to the increased ocean heat.

I think you might be missing the link between ocean and atmospheric heat. The mechanism for the global transfer of heat will likely remain the same--atmospheric circulation, of which hurricanes are a minor, albeit high magnitude element.

To my thinking, with global warming, as rising SSTs occur, the transfer of heat by conduction and convection to the atmosphere will increase, but what will also occur is higher temperatures in the high pressure cells of the subtropical highs. Since water vapor is not an ideal gas, a greater amount of energy can be transferred at higher temperatures, as the mass of water increases with temperature, so there is more energy available via condensation. What this suggests is a higher energy threshold needs to be surpassed (though I am not cocksure, but this would explain the discrepancy in GCMs), but the intensity may be greater--fewer events, but higher magnitude.

As to the methodology for analysis, it appears to be visual graph comparison of some common meterological variables.


I'm a bit off the topic here, but not totally (I don't think). An earlier statement from above said, "After all, we know that storms weaken over cold water, or over land, because they lose their energy source." Okay, makes intuitive sense, but I wonder, particularly with recent news in mind, has there been an observed increase in tornado frequency and intensity, and if so, do the mechanisms and dynamics explaining hurricanes extend to tornadoes, or is there something different going on? It would seem only logical that overall global warming would also have some impact on tornado formation and strength, but I haven't seen any mention of it. Maybe the research just isn't being done, or is at too elementary a level right now. Just curious.

By Harris Contos (not verified) on 16 Mar 2006 #permalink

The causes of tornado and hurricane formations are different. Tornados get their energy from the atmosphere, and usually form when a strong cold front meets a nice, warm, moist atmosphere. Hurricanes get their energy from warm sea water. Of course there are similarities in that condensation releases heat that reinforces the strong upward motion of the air as the storms develop.

mgr, there is a tremendous amoung of energy in warm water. The atmosphere alone doesn't transfer that to cooler regions. Ocean currents are a big part of that. But what happens to ocean currents with global warming is another issue altogether.

By Mark Paris (not verified) on 16 Mar 2006 #permalink

Harris--no, hurricanes and tornadoes are not similar. Tornados are primarily a phenomena of mid latitude North America, while hurricanes are more global. The mechanism, and scaleare different as well. The role that evaporating and condensing water plays in energy transfer is significant to the difference as it is a necessary element to what causes hurricanes, but not tornadoes. No one's certain what the mechanism for a tornado is precisely, but it is associated with cold fronts, and flat terrain. One proposed mechanism (which probably dates me)is that conditions favorable to a tornado formed when the cold front encountered surface friction sufficient to make it 'lag', and with the force behind it, would 'snap' forward abruptly. The explanation never really worked for me, and I have not kept up on it since.



Well done sir.

One thing to consider is not whether the temp difference between sfc and trop narrows - I'm not really sure how that mechanism would occur on the ground - but it is important to understand the lapse rate[1] drives adiabatic cooling as well as temp difference, which is not directly per se affected by absolute trop temp. The mechanism for convection is not as simple as one factor.

Plus, rising SSTs should make for increased evaporation, which helps drive convection, provided the adiabat and lapse rate can keep the drive going. I can't imagine why the atmosphere and surface temp wouldn't rise ~in concert (and certainly not at such a difference to affect convection in any meaningful way - look at RSS MSU analysis).

But I never did tropical wx, so I shouldn't be seen as the last word on the subject.



[1] this is the balloon sounding for Nassau, Bahamas. Lapse rate is how quickly the atmosphere cools and is best visualized in cross-section. The shallower the curve, the steeper the lapse rate [picture it followng the green dashed lines to convect]. Here, there is high pressure aloft negatively affecting convection, which is briefly supported at low levels, but quickly is overwhelmed by the hi press aloft. While this curve follows the green dashed lines, the dry air of high pressure kills upward motion (along with the downward motion of hi of course).

[/lecture mode]

I think one important thing here is the non-linear dependence on saturation water vapor pressure with temperature. If you increase both the surface and the high altitude temperature you will have the same temperature difference, but more energy can be transported by water vapor. An alternative way to state this is that the difference between the dry and the wet adiabatic lapse rate is larger at higher temperature. This mean that you get higher temperature/pressure differences at higher temperature if a packet of air rises along a wet adiabatic, the rain falls out and it then descends along a dry adiabatic.

By Thomas Palm (not verified) on 16 Mar 2006 #permalink

Thanks to you all. You've shown me that I still have a fair amount to learn....

Harris--no, hurricanes and tornadoes are not similar. Tornados are primarily a phenomena of mid latitude North America, while hurricanes are more global. The mechanism, and scaleare different as well. The role that evaporating and condensing water plays in energy transfer is significant to the difference as it is a necessary element to what causes hurricanes, but not tornadoes. No one's certain what the mechanism for a tornado is precisely, but it is associated with cold fronts, and flat terrain.

Tornadoes are probably more of a global phenomenon than hurricnes. They have been observed on all the continents except Antarctica and as far north as 70 N (in Finland). Current estimates are that there are about 300 per year in Europe (not counting waterspouts). The US has ~1200 per year. Southern Brazil and northern Argentina, South Africa, Bangldesh, southeastern China, and Australia also get significant numbers.

The environmental conditions that form the most severe thunderstorms are lots of convective available potential energy [CAPE] (essentially high values of absolute humidity near the ground and steep mid-tropospheric lapse rates) and strong wind shear from the surface to at least the mid-troposphere. The chances go up with strong wind shear in the lowest ~1 km of the environment and low cloud base heights (high relative humidity near the surface). Then, you need some mechanism to initiate storms (front, dry line, outflow from other storms).

Historical changes (since the late-50s) in the US seem to indicate a decrease in the frequency of favorable environmental conditions for severe storms until the early 1970s, followed by an increase of ~<1% per year. There's been no significant trends in the environments favorable for tornadoes.

By Harold Brooks (not verified) on 17 Mar 2006 #permalink

To all those who were kind enough to address my query, thank you, I got a bit of an education there. Not to belabor the point, but if I got it right high humidity in conflict with cold air is a key factor in tornado formation, ergo if global warming -> increased atmospheric humidity, then we would expect to see some effect on tornado formation, although whether (+) or (-) is an open question, if there might be less overall cold air around. Guess it's an area for further research, as are numerous other weather occurrences. One thing's for sure, I won't be able to watch "The Wizard of Oz" the same way ever again.

By Harris Contos (not verified) on 17 Mar 2006 #permalink

Our hero, James Hansen, must be quoted here. Hansen was (and still is) utterly exasperated by the categorical contention of the big cheese at the National Hurricane Center who smilingly said that recent hurricane formation and intensification had nothing to do with global warming. I witnessed that statement first hand and live as I watched CNN and hurricane Rita intesify last fall. During a February 2006 presentation at The New School for Social Research in New York City one of Hansen's slides stated; "I conclude that greenhouse gases are probably responsible for a substantial fraction of the ocean warming that fuels stronger hurricanes."

By gerald spezio (not verified) on 17 Mar 2006 #permalink

Hurricanes often spawn tornadoes when they make landfall. I have no idea how this particualr phenomenon would be affected, though.

More than 90% of the heat from global warming goes into the oceans (with the atmosphere, the two Polar Regions, and mountain glaciers accounting for the rest).

Hurricanes function as a conduit for heat transfere from the oceans to the atmosphere. This is a very important factor in global warming.

Given the tremendous amount of heat soaked up by the oceans because of their ability to absorb sunlight, any efficient exchange of heat between the oceans and the atmosphere bodes ill for the speed of global warming.

In my opinion, non-linear/expotential heat exchange in the form of super-hurricanes are predictable. Due to the tremendous value of ocean front property along the Gulf Coast and the Eastern seaboard, this forebodes a catastrophic readjustment of property values, skyrocketing homeowner insurance rates, and escalating foreclosures leading to a US banking crisis.

Especially worrisome is states and the federal government forming shell corporation as insurers of last resort. In a futile effort to provide affordable house insurance, taxpayers will be stuck with the gigantic costs as increasingly powerful hurricanes hit the coast.

By Brad Arnold (not verified) on 16 May 2006 #permalink