![Awaiting an onslaught: wave activity at Rockaway Beach at midday on Monday, when Hurricane Sandy was 425 miles southeast of New York City.](https://webarchive.library.unt.edu/web/20121211014158im_/http://graphics8.nytimes.com/images/2012/10/29/blogs/waves/waves--blog480.jpg)
4:38 p.m. | Updated with more scientists’ input at the end. }
In the hours before Hurricane Sandy thundered ashore, igniting power stations and Queens neighborhoods and darkening urban and rural communities alike, I filed two pieces for Dot Earth. The first, “The #Frankenstorm in Climate Context,” was on the storm in the context of climate history and the science pointing to greenhouse-driven climate change. The second was on the impact of our tribal nature on climate communication, including my own.
Yesterday, as the winds here in the Hudson Valley rose but before my second piece was posted (and before our power and Internet access were cut as a tree fell down the road), I received an e-mail message from Dan Miller, an engineer and venture capitalist with deep climate concerns, whose name will be familiar to some here because he helped James Hansen, the NASA climate scientist, craft “Game Over for Climate,” a May Op-Ed article in The Times.
Miller strongly disagreed with my discussion of the science revealing past patterns of extreme storminess in the Northeast and the science pointing to a rising human influence on some (but not all) kinds of extreme weather. The exchange is worth posting here now (Miller gave permission), to give you a sense of how intelligent people with related, if not identical, goals can interpret a large body of science very differently:
MILLER:
I just read your piece on Hurricane Sandy, where you make the claim (or at least imply) that it is difficult to attribute the storm to climate change. You point to studies that show that hurricanes have intensified 4 times since the last ice age with a period of about 3000 years.
Your intent is to imply that the current record-breaking storm may be due to natural variation rather than man-made climate change (or at least it’s hard to tell the difference). As you know, “The climate has changed before!” is one of the myths that climate change deniers repeat the most. Of course, the fact that climate change has changed before is, itself, not a myth. However, the implications of the claim that the current changes we are observing are not man-made and/or will not be dangerous are myths.
Do you really think the fact that waters are warmer and atmospheric moisture content is higher now due to man-made global warming (not to mention the “blocking high” over Greenland due to Arctic climate change) may be less of an influence on Hurricane Sandy than some other currently unobserved changes to our climate that occurred 3000 years ago?
I have a prepared a Climate Change Briefing for Policy makers at the request of a congressman which I have attached for your review [available here]. On the last page of the document I address the “But the climate has changed before!” myth. Please let me know if you have any questions, comments, or suggestions about the briefing.
I also refer you to a piece I wrote long ago comparing climate change to the Space Shuttle Challenger disaster. In that piece I point out that NASA managers asked Thiokol engineers to first prove that the shuttle would blow up in order to scrub the launch. Of course, that is the wrong question! They should have asked for assurance that the flight would be safe in order to launch. Your piece on Hurricane Sandy (and others, like your response to Jim Hansen’s Op Ed on the increase in extremely hot summers) are from the point of view of the NASA managers deciding to launch the Challenger. While there is a plethora of evidence suggesting that the path we are on is fraught with peril, you seem to want to focus on the uncertainties in the details of the evidence, rather than taking the prudent course.
When you have the largest Atlantic storm in recorded history that is being feed by unusually warm ocean waters (+5°F) and is being steered in a very unusual direction by a “3-sigma” blocking higher over Greenland after the largest Arctic sea ice melt in human history, you might want to consider the “steroid” hypothesis a bit more.
REVKIN:
First, my only “intent” is to follow the science and assess where it leads in terms of policy and personal responses.
I’d have to point out that you’re missing the point of my piece, and the significance of “Millennial-scalestorminess variability in the northeastern United States during the Holoceneepoch,” the 2002 paper, coauthored by Realclimate’s Eric Steig, that I focus on.
The paper created a Holocene-length record of extreme precipitation, as derived from lakebed sediments across the Northeast. So it captured extreme inland rainfall from tropical systems (as in this system, Irene, Floyd, and Vermont’s epic 1927 gullywasher) as well as nor-easters and any other rare events. Second, the authors predict we’re in an era of rising storminess of that kind in the Northeast, via ocean cycles. There’s more in the paper conclusion below.
So it’s fine to talk about sea level and surge risk from hurricanes, but not fine to describe the weirdness factor in this hybrid system as somehow significantly shaped (emphasis on “significantly”) by greenhouse forcing. And that’s what some are doing.
Here’s the conclusion:
Climate models suggest that human activities, specifically the emission of atmospheric greenhouse gases, may lead to increases in the frequency of severe storms in certain regions of the Northern Hemisphere. However, the existence of natural variability in storminess confounds reliable detection of anthropogenic effects.
During the past ~600 years, New England storminess appears to have been increasing naturally. This rhythm in storm frequency may explain some of the recently observed increases in extreme precipitation events. If the pattern of millennial-scale variability that we documented through the Holocene persists into the future, New England storminess would continue to increase for the next ~900 years. Because climate synopses compiled from instrumental records cannot distinguish underlying natural increases in storminess from anthropogenic effects, detected increases in contemporary storminess may not be a reliable indicator of human-induced climate change.
Miller:
Your response…follows the same line of thinking that I was addressing in my email to you. Yes, there has been more and less “storminess” in the past. Yes, we may even be in for a “natural” increase in storminess over the next several hundred years. But mentioning that in the context of Hurricane Sandy is an attempt (or at least has the effect) to minimize the concern about the documented changes to the climate we have already made and the predicted catastrophic consequences of staying on the fossil emissions path we are on. It is seeding doubt at a time when action is needed.
We have increased CO2 levels in the atmosphere by about 40% in the last 100 years (mostly the last 50 years) on the way to doubling later this century. The Earth has warmed up about 0.8°C (1.4°F) already due to the extra greenhouse gases we put in the atmosphere and it would have warmed even more if we weren’t also putting up smoke that reflects sunlight. This warming has increased Earth’s energy radiation to space, but the excess greenhouse gases are still trapping more heat than the Earth is radiating to space. This “energy imbalance” is about 0.6 watts/square meter. This doesn’t sound like much but it is equivalent to 400,000 Hiroshima atomic bombs going off every day (see my briefing for the math).
Extremely Hot Summers (“3-sigma” events) have increased 50X (5000%) in the past 50 years. There is 4% more water vapor in the atmosphere than 50 years ago. Average ocean temperatures have increased (90% of global warming energy goes into the ocean). The Arctic sea ice just reached its lowest level in thousands of years and in a few years you will be able to sail a boat to the North Pole for the first time in human history.
These documented impacts all effect the strength, scale, and direction of Hurricane Sandy. No one is saying that a Hurricane Sandy would not have happened if not for climate change. But I believe there is little doubt that the record-breaking scale and potential destructiveness of Sandy is due in large part to the amplifying effects of warmer ocean temperatures, higher atmospheric moisture content, and unusual Arctic weather patterns.
Like the Space Shuttle Challenger’s NASA managers, waiting for scientific “proof” of disaster, rather than taking prudent (and economical beneficial) steps to avert disaster, only guarantees that our children will face catastrophic consequences.
Some of the points Miller makes about Arctic influences on Hurricane Sandy and the resulting nor’easter are explored by Kevin Trenberth and other scientists in my post on the storm and climate science. Have a look. Also check yesterday’s post on late-season hurricanes and global warming by Eric Berger of the Houston Chronicle for more on the science.
And of course I’ve never bought into the “wait until there’s proof” approach to human-driven climate change, which — as I’ve stressed since 2007 — is more a challenge of risk management amid substantial and persistent uncertainty than a problem resolved by amassing scientific proof.
I still like how Carl Wunsch of the Massachusetts Institute of Technology summarized things in my 2007 article, “A New Middle Stance Emerges in Debate over Climate” (of course that article prompted lots of slings and arrows, too):
“Climate change presents a very real risk,” said Carl Wunsch, a climate and oceans expert at the Massachusetts Institute of Technology. “It seems worth a very large premium to insure ourselves against the most catastrophic scenarios. Denying the risk seems utterly stupid. Claiming we can calculate the probabilities with any degree of skill seems equally stupid.”
If you could enter this discussion, what would you ask either of us? I’ll alert Miller as questions come in (and I’ll answer those directed at me).
4:38 p.m. | Updated | I read Mark Fischetti’s piece on global warming and hurricanes in Scientific American just now, which points to a recent PNAS study finding “a statistically significant trend in the frequency of large surge events” from tropical cyclones in the Atlantic.
The same paper was discussed in my Sunday post, and when I reached out to other scientists, it didn’t get a very promising reception. Both Tom Knutson and Gabriel Vecchi, veteran hurricane-climate researchers for the National Oceanic and Atmospheric Administration, offered detailed critiques that I added as comments on Sunday. Comments are easy to miss, so here are excerpts and links:
Knutson:
On inspection, I am cautious about making too much of this result, as it appears to be a case of marginal significance for several reasons:
1) The start date in 1923 is a relatively low period in landfalling and basin-wide hurricane/or tropical storm time series with relatively higher activity in the late 1800s. This can be seen for example in Fig. 3 of this web page that Gabe Vecchi and I have.
Our experience is that starting a trend in a relatively low point of a slowly fluctuating time series can sometimes lead to findings of statistical significance which don’t hold up as the record is further extended (e.g. extended further back in time or new years added on the end). [Read the rest.]
Vecchi:
This is an interesting paper, but I share the concerns that Tom has raised. In addition, although they have shown a correlation between their surge index and measures of hurricane activity, it is far from a 1:1 relationship (correlations of 0.5-0.6 indicate that this index shares about 25-36% of its variance with direct hurricane measures, leaving over 60% to be explained). As can be seen in the plot if the two indices in 2005, the biggest storm surge indices do not match the biggest ACE events (and vice-versa). That makes sense, since there are many factors that impact storm surge, of which intensity is only one. For example, the trajectory, speed and size of the storm all influence storm surge. Therefore, one needs to be wary about over-interpreting these results. Peak surge has shown a trend, but what does that tell us about storms themselves? What is behind this trend?
I find the correlation to SST a bit of a non-sequitur given that the factors responsible for the surge trend are not clear. Is this sea level rise? Stronger storms? Storms with different tracks? Starting in 1923 adds to the problems in interpreting the trends. [Read the rest.]