Katharine Hayhoe on the Sandy/Climate connection

Katharine Hayhoe on Twitter (or rather the service TwitLonger since 140 characters is obviously not enough) explaining climate change’s influence on Sandy, and the difficulties in communicating the nuances involved to the press:

As a scientist in a field where consensus is constantly challenged, I’m a bit sensitive about being pitted against a colleague. Case in point, a recent USA Today article, which includes the following quotes:

“Mostly, it’s a natural thing,” climate scientist Gerald North of Texas A&M called the storm in an interview with the Associated Press. Others, such as Texas Tech climate scientist Katharine Hayhoe, noted that warmer Atlantic waters and Arctic weather that helped steer the storm were juiced by global warming and played a significant role.

Full Story: http://usat.ly/SJ7PaP

Here’s the funny thing: I actually agree with Gerry (a well-respected colleague and one who is just as concerned about the potential impacts of climate change as I am). Yup, I do. What do I mean by that?

First, large hurricanes in October are not at all unusual. About one of them hits the US every 5 years. This time, it was Sandy.

So did climate change cause Sandy? No: there is absolutely no evidence that it did. In fact, several future modeling studies suggest climate change may actually *decrease* hurricane frequency (although increasing the number of Cat 4 and 5 storms).

Second, while climate change more than likely exacerbated the impacts of the storm, the influence of climate change on sea level and ocean temperature is clearly *less than* natural variability at this time.

I estimated earlier (see previous twitter feed for references) that long-term 7 inch global sea level rise could have enhanced Sandy’s storm surge by 4% and sea surface temperature increases by something on the order of 1oC. Compare that to the 14 foot storm surge measured during the storm, and the >5oC natural SST temperature anomaly present this month.

So Gerry’s quote is entire consistent with both my and other mainline scientific understanding of this issue: the data does suggest most of Sandy was natural. But I would bet that Gerry would also agree with me, that it’s not likely that the impact of climate change on Sandy was zero.

For now, though, we have absolutely no way to quantify or put a number on the impact of climate change on Sandy. We do know it wasn’t zero, and we do know it wasn’t 100% and probably not even 50% (which would qualify as “most”).

Now, in between 0 and 100% is a big range and you can bet that many scientists will be working on narrowing down that gap in the next few years. Stay tuned to the glacial pace of scientific journals over the next three years to find out more :)

Comments:

  1. This is pretty much wrong in my opinion. Most of it strikes me as wrong-headed in that Marty Hoerling way. It's a fairly involved argument but I think it's an unworkable approach to attribution in a nonlinear system.

    Furthermore as far as I know "the influence of climate change on sea level and ocean temperature is clearly *less than* natural variability" is flatly wrong on both counts for any reasonable interpretation of those words.

    She's been hanging around Lubbock too long, I reckon.

    • Is sea level rise more than natural variability? I think it depends on the timescales you want to consider, if you look at daily variability (tides) then clearly natural variability is higher. The same thing goes for sea surface temperature (though on seasonal instead of daily time scales).

      • Of course as soon as I posted the above comment I realized the problem.

        We don't say warming is still within natural variability because the difference in temperature from day to night (or summer to winter) is larger than the increase in the warming anomaly, so I can see why we doing the same thing for sea level could be problematic.

        Yet at the same time when talking about the hurricane the fact that it coincided with a spring tide probably contributed more to the devastation than the change in sea level due to climate change.

  2. the attribution of causation is always problematic with any system that cannot be reduced to a singular factor with a binary effect. Most people do not deal well with multiple factors with percentage influences. Even worse is the grasp of attribution in a nonlinear system.
    Its probably part of the deficit theory-{grin}.

    But thresholds are a simple enough concept for most people to grasp. So just for the sake of rhetorical argument lets take the idea that Sandy was mostly Natural with only a <50% kick from AGW.
    Lets attribute 20% of Sandy to AGW. all of the influence act to intensify the storm. The wobble of the jet stream may be uncertain in its attribution to AGW, but a 20% greater chance of hitting the coast instead of drifting east seems appropriate. Then shave 20% of the wind speed, storm surge and extent off Sandy. You still have a pretty savage storm, but it would cause much less than damage than Sandy which is Natural plus 20% AGW.

    The extra 20% has a lot more effect than the first 80% when it comes to the damage inflicted. It can make the difference between a sea defense being sufficient for a surge under 12ft, but overwhelmed by a 14ft surge.

    Nature can be extreme and savage without any push from AGW, but with a Lance Armstrong climate it has the little extra to make major events exceptional.

  3. MT, I think I understand your objection to Hayhoe's attribution method (the +1 deg* probably has a disproportionate effect) but why is her statement about natural variability wrong? Tides still dwarf sea-level rise to date. Ditto SST variability, I think.

    (*She said that CC is responsible for about 1 deg C of a 5 deg C SST anomaly. Trenbirth said 1 deg F of 5 deg F. Who is right? Him, her, both, neither?)

    • I do think Katharine Hayhoe was referring to tides and seasonal changes in sea surface temperature when she said that natural variability dwarfs the effects of climate change. As I mentioned above, I think the potential problem with comparing sea level rise with tides is the same as comparing the warming anomaly with the change in temperature from day to night or summer to winter.

      But the devastation caused by Sandy had a lot to do with the tides, so I don't think it the correct way to communicate this is obvious.

    • Neither of them. I am (perhaps stupidly, surely quixotically) willing to go against Kevin on this one.

      Reading Kathy's post I see what she meant, but I think it is misleading. As for the 1 degree of 5, that's a lower bound, not a best estimate. It just doesn't work that way. This is the same mistake people make when they attribute only a tiny fraction of last year's Texas heatwave to what everybody insists on calling "global warming". It's not a linear superposition.

      And even in linear systems, large transient forcings normally increase system variance. I have yet to see any reasoning why the climate system should behave otherwise.

      Hansen recently showed clear evidence that in fact some pieces of the climate system are moving into higher variance. Other parts may well be, but the change may not yet have emerged from the noise, or the right analysis might not yet have been discovered to identify it.

      I do not know how uncommon a 5C anomaly in sea surface temperature off New England is. If it is common, then my argument is moot. But if it is rare, I disagree with what you are attributing to Kevin. In general, we are getting too many rare circumstances.

  4. Kathy Hayhoe has contacted me with this link. I hope we can get her to engage further.

    I am strongly inclined to believe that there are footprints of climate change all over the Sandy episode.

    I am not sure this is demonstrable in any straightforward way, but I think arguments like Kathy's, while safe and hard to argue against, greatly understate the case.

    That said, I suppose one should admit that the phase of the moon is not anthropogenic.

    • Yeah, several footprints. I (nonexpert) see two small ones and a big one. Perhaps there's more (e.g. tropical belt expansion)?
      * Sea level (and 20cm atop high tide isn't trivial, alas not enough to wash Wall Street)
      * SST
      * New arctic meteorology (on Neven's blog they saw weirdness coming).

      That's why mestupid thinks Sandy is quite obviously "Gaia's Revenge" (for her "Vanishing Face", i.e. arctic sea ice. James Lovelock's book titles.) No detailed science and modelling necessary to accept this.

    • P.S.: I love the "footprints of climate change" phrase. Footprints is what Sherlock Holmes (i.e. Joe Sixpack climate critic) is looking for first.

  5. Hi guys,

    The main point of what I wrote above was to express my frustration at how the media will still pit two scientists against each other even when they actually agree, the article on the hurricane being a case in point. Yes, it makes a better story, but doing so just reinforces once more the myth that scientists are still arguing about climate change. (I should note that the writer of this piece very kindly acknowledged my objection and has promised to change it!)

    Because my point was to show that Jerry and I agree, and because I had already said quite a lot about the relationship between climate and hurricanes earlier both on twitter as well as here (http://stateimpact.npr.org/texas/2012/10/30/three-ways-climate-change-made-hurricane-sandy-worse/), in what I wrote above, I focused primarily on the natural contribution to climate change above. Without the context, it is hard to get the full picture.

    That said, I stand by what I said: there is no way to quantitatively estimate the percentage contribution of climate change to hurricane Sandy without several years of very detailed modeling that includes not only the physical impacts on the hurricane itself, but the relationship between hurricane strength, pathway, and impacts. Not a trivial task. (You might be interested to know that I disagree with Marty on this point: he says there is no way. I say there is, theoretically at least; but it is not easy!). We know it's not zero, but we know it's not 100%. We can try to narrow that gap by piecing together what information we currently have available, but we have to be honest and call it what it is: guessing (some more, and some less, informed!).

    Katharine

    PS> Vinny, I put all the data I used and figures on twitter. I try to source all the statements I make, in this case including the ones about slr, contribution of long-term change to October temp anomaly off the east coast, etc. You can check there if you want more info.

    • KH, your long Twitter piece (linked above) said something like 'see previous tweet for the references'. I went looking for them but no luck. Any chance of a direct link to the tweet in question? Please.

    • Thanks for your participation, Katherine.

      You said:

      "That said, I stand by what I said: there is no way to quantitatively estimate the percentage contribution of climate change to hurricane Sandy without several years of very detailed modeling that includes not only the physical impacts on the hurricane itself, but the relationship between hurricane strength, pathway, and impacts. "

      I completely agree with that. It would be hard to find a scientist who doesn't.

      What I took issue with was the statement that "while climate change more than likely exacerbated the impacts of the storm, the influence of climate change on sea level and ocean temperature is clearly *less than* natural variability at this time." I see now that you were comparing sea level rise with the tidal signal. This is wrong to the extent that the storm surge lasts a long time compared to 12 hours and is thus likely to be near enough peak at high tide. You should be comparing the sea level to the lunar tidal cycle, (the difference between spring and neap tides) which is comparable.

      And as for ocean temperature, I am not convinced the ocean will behave differently under greenhouse transients than the atmosphere, increasing the likelihood of outlier events more than would be indicated by the shift in the mean as in Hansen's recent work. So that would mean more than 1/5 attribution there.

      Also, we have the very large storm extent and the very deep trough, both of which are clear in current trends. Though the mechanism connecting them to greenhouse forcing is unclear in both cases, it would have to be the prime suspect.

      I do not think it is unreasonable to look at Sandy as the shape of things to come.

      There is a silver lining specifically regarding tropical storms. The size of the storm increased the fetch and hence the surge, but it decreased the pressure gradient and hence the central winds. Indeed, it seems to me that recent hurricanes have had far less wind damage than surge and flood damage.

  6. I used Surging Seas data to quantify how many more people were affected by the increased sea levels in NY City. Global warming made Sandy worse.

    71,000 New Yorkers: “Rise Does Matter!”
    http://profmandia.wordpress.com/2012/11/05/71000-new-yorkers-rise-does-matter/

  7. Pingback: Another Week of GW News, November 4, 2012 – A Few Things Ill Considered

  8. The planet's circulatory system is like the human body, not a house built of bricks. You can't cut off an arm and think about the rest. All the pieces are integrated with each other.

    Scientists (and men) want clarity. But the planetary systems are complex and beyond containment in a box. They are not a computer, they are real.

    I find the ongoing need to find a baseline of quantifiable very irritating. We should be far more open to describing and understanding what we can and accepting whatever information is available about the rest. That ambiguity is built in to being human, and if we could all stop playing roulette with reality because we cannot make everything finite and describable, we could get on.

    BTW, I'm with Dr. Hayhoe, and the rest of you too, except when your faultfinding takes over. It's a kind of bias, and we can't afford it.

    • I am *not* against quantifying, computers, or working to gain understanding. Sorry if I didn't make that clear. I'm for integrating what we can understand with what we can and have observed and working with what we cannot understand in a way that doesn't refuse to work with what we cannot measure.

      • http://www.dailymail.co.uk/sciencetech/article-2217286/Global-warming-stopped-16-years-ago-reveals-Met-Office-report-quietly-released--chart-prove-it.html

        Data quietly released by MET. Shows no warming in 16 years.

  9. A little late to the party here, but I just want to point out that 7" of water above the floor level of a house is notably different than 7" of water not above the floor level of a house.

    Going forward, we'll always be able to say that the past few years of sea level rise are insignificant when compared to a given storm surge, but 4% of 14' of water is 100% significant when it has invaded your home.

    If you've had a house with just 1" of water sitting above floor level, you'll know what I mean.

    Here's our house in Seattle, with "just" 9" of water in the downstairs:

    http://www.skepticalscience.com/pics/water-not-good.jpg

    The house just isn't as useful as it was without the water inside it. In fact, just the smallest amount of water inside renders a home practically unusable, particularly if soft furnishings are involved. In our case the drywall you see was installed just 2 days prior to the arrival of this insignificant amount of water, the house being in the middle of a remodel, so we were "lucky." A mere $11,000 got us all fixed up.

    That water arrived as part of historic record rainfall for the Seattle area, 6" at the nearest NWS station to us. How much of it was climate change? An inch? Once the first inch of water enters the house, who cares? Life is rearranged.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>