02 November 2012

A Summary of Sandy Discussions

Here is a short guide to the various discussions of Sandy on this blog and from a few of my Tweets this week.


  1. I just entered this at the Tamino blogon this thread:


    Reproduced here just in case ;-)

    "Your comment is awaiting moderation.

    It is great to see interest in the science of disasters, fortunately there is a growing number of empirical studies of this subject. Here are a few:

    recently, Munich Re provided a large grant to LSE to examine the reasons for the trends in its catastrophe database. The resulting analysis was published in GEC last year. Here is what they found:

    “Independently of the method used,we find no significant upward trend in normalized disaster loss.This holds true whether we include all disasters or take out the ones unlikely to be affected by a changing climate. It also holds true if we step away from a global analysis and look at specific regions or step away from pooling all disaster types and look at specific types of disasters instead or combine these two sets of dis-aggregated analysis. Much caution is required in correctly interpreting these findings. What the results tell us is that, based on historical data, there is no evidence so far that climate change has increased the normalized economic loss from natural disasters.”

    E. Neumayer and F. Barthel, Normalizing economic loss from natural disasters: A global analysis, Global Environmental Change, 2011, ISSN 0959-3780, DOI: 10.1016/j.gloenvcha.2010.10.004.

    This Munich Re-funded study is of course consistent with the broader peer-reviewed literature, recently summarized by L. Bouwer in BAMS:

    “Economic losses from various weather related natural hazards, such as storms, tropical cyclones, floods, and small-scale weather events such as wildfires and hailstorms, have increased around the globe. The studies show no trends in losses, corrected for changes (increases) in population and capital at risk, that could be attributed to anthropogenic climate change. Therefore it can be concluded that anthropogenic climate change so far has not had a significant impact on losses from natural disasters.”

    Bouwer, L. M., 2011: Have Disaster Losses Increased Due to Anthropogenic Climate Change?. Bull. Amer. Meteor. Soc., 92, 39–46.
    doi: http://dx.doi.org/10.1175/2010BAMS3092.1

    The peer-reviewed literature on this topic was recently summarized also by the IPCC SREX which concluded:

    “Long-term trends in economic disaster losses adjusted for wealth and population increases have not been attributed to climate change, but a role for climate change has not been excluded (medium evidence, high agreement). . . The statement about the absence of trends in impacts attributable to natural or anthropogenic climate change holds for tropical and extratropical storms and tornados. . .The absence of an attributable climate change signal in losses also holds for flood losses”


  2. Roger -

    ==]] The studies show no trends in losses, corrected for changes (increases) in population and capital at risk, that could be attributed to anthropogenic climate change. [[==


    ==]] The studies show no trends in losses, corrected for changes (increases) in population and capital at risk, [[==

    Was there also control for infrastructure development, advanced methodologies for responding to severe weather events, and any possible gains due to better forecasting, better understanding of severe weather events patterns? Obviously, the impact of those factors on economic damage might be limited - but seems like it should be considered as caveats and reflected in qualifications of certainty.

    Also, it seems that some consideration should be given to the ratio of storms/storms that make landfall. Probably unlikely (especially on a global basis), but it is theoretically possible that there is a lower % of storms making landfall due to a temporary artifact of climate change.

  3. Tamino offers this by way of response:

    "[Response: Odd ... this post isn't about economic losses but about the number of disasters, yet you've said not one word on that subject.

    I'm hardly surprised that you changed the subject, because I count you not as a skeptic but as a denier. This is not a forum for denier propaganda. I suggest you use your own blog.]"

    I then submitted this reply:

    "Thanks for the reply.

    The graph that you show at the top of this post is in fact a count of "catastrophes" which are defined by Munich Re as disasters which exceed a certain economic threshold of loss. So my comment is in fact directly responsive to the topic that you have raised.

    As far as the name-calling, you can call me whatever you'd like, but perhaps we might choose to focus on the science instead?


  4. Tamino has changed his response to me as follows:

    "[Response: How interesting that this post isn't about economic losses but about the number of disasters, yet you've said not one single word about that. I'm hardly surprised that you changed the subject, because I count you not as a skeptic but as a denier.

    This is not a forum for denier propaganda. Should I ever post specifically about your statements or publications you will be free to respond unhindered. Until then, you will not hijack my threads. I suggest you use your own blog.]"

  5. Sorry - cut-and-paste problems. The second excerpt related to control of data should have been this:

    ==]] "...adjusted for wealth and population increases " [[==

  6. Tamino responds with the class and dignity that I've come to expect from the "climate change people":

    "[Response: Perhaps I wasn't clear enough. I'm happy to discuss the science with those who are able to do so honestly. I do not count you among them.

    As for the term "denier," I will not refrain from calling you what you really are. If the term offends you, somehow that news fails to disquiet me.]"

    I offered this further response:

    "Your comment is awaiting moderation.

    Thanks for your further response. Again, I really don’t mind if you think I’m dishonest or if you’d like to call me names, that is your prerogative, no worries.

    However, whatever your views about me, it is hard to connect that to your characterization of the peer reviewed research I have cited (one study funded by Munich Re to look at the dataset you posted at the top) as “denier propaganda.”

    Rather than engaging me, perhaps you might engage the actual science related to the Munich Re catastrophe dataset found in the peer reviewed literature? Here is a hint to better understanding the trends you see — the threshold for inclusion on the list is a function of economic losses and is not a constant value.


  7. Roger -

    Just curious if you have any thoughts about calling people "climate chickens?"

    Classy? Dignified? Is that what we should come to expect from people who interpret the science as you do?

    It seems contradictory to, on the one hand, say that you unconcerned about being called a denier, and to then imply that someone doing so lacks class and dignity - particularly since you engage in the same sort of behavior.

  8. Tamino tells me to go away:

    "And now, since you seem too dense to get the point, I will be explicit: you are not welcome here."

    I'll respect his wishes ;-)

  9. Roger - since you've been dis-invited participate at Tamino's, I'll reproduce an interesting comment from over there (minus the comments related to the non-science components of the exchanages):

    MapleLeaf | November 3, 2012 at 9:10 pm | Reply


    Odd that Roger forgot to mention the other finding by Neuymayer and Barthel (2011), namely:

    “Due to our inability to control for defensive mitigation measures, one cannot infer from our analysis that there have definitely not been more frequent and/or more intensive weather-related natural hazards over the study period already. Moreover, it may still be far too early to detect a trend if human-induced climate change has only just started and will gain momentum over time.”


    Barthel and Neumayer (2012),

    “We find no significant trends at the global level, but we detect statistically significant upward trends in normalized insured losses from all non-geophysical disasters as well as from certain specific disaster types in the United States and West Germany.”

    Schmidt et al. (2010) did find evidence of an increase in adjusted losses from climate change,
    “The findings show the increase in losses due to socio-economic changes to have been approximately three times greater than that due to climate-induced changes.”


    Elsner et al. (2012), summarized here by Elsner on his webpage:

    “While the SSTs did not proximally cause Sandy to curve into New Jersey, they quite likely caused Sandy to be stronger. Our new research shows that the limiting intensity of hurricanes (how strong hurricanes can get as a statistical limit) relates to SST at about 8 m/s/C. With SSTs in the path of Sandy that were 2-3 C warmer than is typical, we predict strong hurricanes like Sandy to be twice as strong on average. Obviously a 45 mph Sandy at landfall would have been considerably less destructive.”

    Additionally, and Murane and Elsner (2012) find,
    “Here we show that the relationship between wind speed and loss is exponential and that loss increases with wind speed at a rate of 5% per m s−1″,


    “The exponential relationship suggests that increased wind speeds will produce significantly higher losses; however, increases in exposed property and population are expected to be a more important factor for near future losses.”

    That is just one aspect, we have not addressed flooding from increased rainfall from global warming, or the increase in storm surges (e.g., Grinsted et al. 2012) associated with warming.


    In particular, I'd be curious to hear your take on the comments about lack of control for variables related to defensive measures and the non-linear relationship between growth in wind speed (and I would imagine also surge) and damage.

  10. -9-Joshua

    Thanks for the questions about this area of research a few replies will follow in the next few responses:

    1. "defensive mitigation measures"

    Any bias in a loss normalization can be detected by comparing (a) the loss normalization trends, and (b) the related trends in the relevant geophysical phenomena.

    If there is an unaddressed bias, then the trends in (a) and (b) will diverge.

    So for instance, take a look at this post on hurricane normalizations:


    With no trends in hurricane frequency or intensity, on what basis would one expect to see trends in a loss normalization? Our new tornado paper performs a similar test, and things match up very well. For floods there are no trends in extreme streamflow, but flood losses/GDP have gone down.

    A further point is that in such arguments those raising them seem to think that the effect of societal change is to reduce losses over time. That is not well supported empirically. In the US the building stock that performs best in hurricanes is from the first half of the 20th century, see:

    Of course, as we have written many times, if your goal is to detect trends in climate phenomena, then it is best to look directly at actual climate data.

  11. continued...

    2. " it may still be far too early to detect a trend if human-induced climate change has only just started and will gain momentum over time"

    Very true.

    We have addressed signal detection in this paper:

    Crompton, RP, RA Pielke and KJ McAneney (2011), Emergence timescales for detection of anthropogenic climate change in US tropical cyclone loss data. Environ. Res. Lett. 6 (1) , doi: 10.1088/1748-9326/6/1/014003.


  12. continued...

    3. "from certain specific disaster types in the United States and West Germany"

    True. So this means what?

    It is a big planet, I can find up and down trends for various phenomena over various time periods. Such arguments are very similar to "Temperatures have cooled in Atlanta over 80 years so global warming is false."

  13. continued...

    4. "Schmidt et al. (2010) did find evidence of an increase in adjusted losses from climate change"


    From their paper: "There is no evidence yet of any trend in tropical cyclone losses that can be attributed directly to anthropogenic climate change."

    Schmidt et al. almost perfectly replicate our loss normalization using Munich Re data, there is no long-term trend:


  14. continued...

    5. Elsner et al. (2012)

    I don't see how this is relevant to the Tamino post or normalization research.

  15. continued...

    6. "Murane and Elsner (2012) find..."

    I found this first;-)

    Pielke Jr R. A. 2007. Future economic damage from tropical cyclones: Sensitivities to societal and climate changes. Philosophical Transactions of the Royal Society A, 365:2717–2729.

    Glad to have the results confirmed (and independently Bouwer also finds something similar in his response to Nordhaus, a paper that also predates Murnane and Elsner).

  16. I believe that is it. Follow ups welcomed.


  17. Thanks Roger -

    I'll look in more detail later.

    One more question - given the anomalous degree of impact from the 1926 storm, I have to wonder what your graphs would look like if that storm were treated as an outlier. My guess is they wouldn't be significantly affected - but I do wonder what the statistical practices should be for treating something that is so far off the scale of the rest of the data.

  18. You appear to have missed Grinsted, et al, who show that
    The surge index is positively correlated with all of the comparison measures. The best correlations are found with measures that emphasize intensity (e.g., NTC, ACE, and PDI) and measures that are restricted to US land-falling storms only. Table 1 also shows that low-frequency correlation tends to be at a higher level than the year-to-year correlation. This is to be expected given that there are low-frequency driving agents related to various climate forcings (4, 8). One notable exception is NHD, which shows poor low-frequency correlations.

    That means that the LONG TERM correlations of NHD with physical measures of hurricane damage will be lousy, but short term correlations will be high (or as high as anything else). NHD in that sense is a measure of the value and density of structures at the time that a hurricane hits, but it is lousy at capturing long term improvements.

    Also, the entire insurance industry is busy fleeing from your opus. The rats know when a ship is sinking.

  19. -18-

    Thanks ... In due course I'll have a post on Grinsted et al., I have a request in for their underlying data.

    It will be interesting to see how they get an increase in surges from hurricanes, yet there are no increases in the frequency or intensity of landfalling storms. Unfortunately, there is a lot of non-hurricane data in their time series. But I'll report back once they provide the data.

    You are incorrect when you write "That means that the LONG TERM correlations of NHD with physical measures of hurricane damage will be lousy"

    The correlation of landfalling storms and damage is 0.71 since 1900.

    "the entire insurance industry is busy fleeing from your opus"

    Thanks for the tip;-)

  20. Roger -

    Thanks again for the responses.

    You are normalizing data going back to 1900 and extending to 2012, but your link about the impact of building practices refers to comparing building practices from the '70s to 90's to building practices post 1926 at the earliest and only up to the 50s (approximately 1/4 of the time period you are normalizing).

    And as far as I can tell, it only refers to residential construction.

    Consider, for example, all the buildings constructed in the late 60s that were potentially better built than those constructed in the 1910s or 1940's that were still standing when a hurricane hit in the 80's or 90's. How have you discounted for those sorts of factors when you eliminate the potential of better construction to minimize loss in later years?

    I get the basically robust methodology of comparing trends to identify biases. I still think, however, that there are way too many uncontrolled variables to distinguish causality to the extent that you seem to be doing.

    I have to admit a bias - having worked in construction, my experience is that when people say "They don't build them like they used to" they are often wrong. Older buildings were overbuilt in many ways, but I've also seen many older buildings where you'd have to just stand back and marvel how they were still standing.

    That isn't to say that I don't think that the data are interesting and important. In particular, I think that they can serve to as a hedge against facile conclusions of causality that go in the other direction.

  21. Roger - I posted a comment about the incongruity between the period of time for your normalized data on storms and the data you used to examine the impact of changes in building standards. Is there some reason it went into your trash or did it just get lost?

  22. -21-Joshua

    Sorry, but there are no comments uncleared or in spam. Please resubmit. Thanks!

  23. So Roger - Any response to explain what your link about building stock (relevant to a time period of 1/4 of your data) was supposed to show, and no response on whether the 1926 storm is a statistical outlier?

  24. -23-Joshua

    Thanks ...

    1. Please read the NHR paper, the point is that construction practices do not improve with time

    2. Here is our data, have at it:


  25. Just a quick question about your summary of the blog post: "Does it make sense to use energy policy to modulate future disasters? No." Does that point referring to all disasters or just storms? Or is that storms are the major identified disaster threat from warming. I am asking as a neophyte, so I have no agenda.