16 May 2012

Updated: The US Intense Hurricane Drought

In preparation for an upcoming talk, I have updated the figure above to the start of the 2012 hurricane season, which will begin with a record-long stretch of no intense hurricane landfalls still continuing. (In most browsers you can click on the figure for a larger view.) The long stretch with no intense hurricane landfalls has surely shaped expectations, setting the stage for all sorts of animal spirits to be in play. Oh, to be a commodities trader this summer.

Below is a figure showing the intra-seasonal distribution of US intense hurricane landfalls. About 8% have occurred in June and July, and more than 75% in August in September. (UPDATE: This figure has been updated from the original posted to fix an error in the median date.)

21 comments:

  1. .
    Clearly this catastrophic lack of big hurricanes is a threat to all mankind. If this news gets wide circulation, those brave souls that are trying to save the planet from the danger of exhaled gasses will have a far harder time saving us all.

    .

    ReplyDelete
  2. There's a technical term in the recent climate science literature to describe the conditiones observed. What was it again? Oh yes, "unprecedented".

    I do give Big Alarmism credit for recognizing and exploiting the equally extreme low numbers between storms that the 2005 storm season provided. But that one inconvenient data point of 2,412 days kind of monkey-wrenches their elegant theoriy, does it not?

    ReplyDelete
  3. I'll bet that with the right smoothing filter, you could make that into a hockey stick!

    ReplyDelete
  4. I remember a few years ago, after Katrina, the AGW mob was predicting more violent hurricanes every year in our near future...

    ReplyDelete
  5. The chaotic nature of the system has not realized the necessary confluence of events to exhibit any significant anomalies.

    Dr. Pielke, with an incomplete characterization of the system, and an unwieldy composition, how reasonable is to estimate its statistics outside some finite period? Why isn't a chaotic model employed with the accuracy acknowledged to be limited to its behavioral envelope? It seems that would be sufficient for planning purposes.

    ReplyDelete
  6. PROF PEILKE

    well I am grain and energy Meteorologist so this Hurricane season is going to be a zoo no matter what. Undoubtedly the argument will be made by some clown somewhere this August September or October that the 6 year severe Hurricane drought is somehow connected to AGW.

    Interestingly the same sort of thing was being said and argued after the severe Hurricane seasons of 2004 and 2005

    DT
    wxrisk.com
    Richmond VA

    ReplyDelete
  7. This may be the silliest thing I've ever seen. The timing of major hurricanes making U.S. landfalls is independent of the previous landfall. There's a equal probability of a major hurricane making landfall from year to year. The cumulative probability density function says nothing about the independent probability of a major hurricane making landfall in the U.S., only that most major hurricanes occur from late August through Mid-October. So, for the sake of the commodity trader, they should stick with their weekly to monthly models and consulting meteorologists and not get too excited by these seasonal statistics.

    ReplyDelete
    Replies
    1. I would disagree. Since these cycles are clearly multiyear or even decadal, a year shown to be a few years into a multiyear low is clearly more likely to be a low year and the opposite is true.

      Delete
  8. Who among you can tell me why the median date (September 8) of intense hurricane landfalls in the US is so very believable?

    ReplyDelete
  9. This may be the silliest thing I've ever seen. The timing of major hurricanes making U.S. landfalls is independent of the previous landfall. There's a equal probability of a major hurricane making landfall from year to year.

    You think?

    It is however the reverse of the alarmist allegation that the frequency of serious hurricanes was rising. That means that the probability of hurricanes is not equal, but time dependent.

    Your allegations of stupidity should be thrown at those making the claim that the rate of hurricane landfalls is changing. Not Roger.

    ReplyDelete
  10. Sorry, Professor. I visited your father's site first. Congratulations, Dr.

    ReplyDelete
  11. Poster 8. Brian.. I would have to agree with Mark take the catastrophe modeling companies. They have specifically come up with stochastic catalogs that increase the landfall due to the warming of the sea surface temperatures. It is these new catalogs that were developed after the 2004, 2005 season that have helped drive up the coast of insurance and reinsurance. The question is after the cold AMO sets in will they release a new catalog that would decrease the number of landfall? The U.S. historical landfall recorded is one of the best natural hazard catalogs we have and making assumption that warmer seas surface temperatures would increase landfall has failed. I also agree stick to the seasonal forecast models, ENSO has a far greater connection to landfalls than the overall sea surface temperature.

    ReplyDelete
  12. @WestHoustonGeo:
    http://en.wikipedia.org/wiki/Galveston_Hurricane_of_1900

    I spent many happy hours in the Bishop's Palace and the Moody Home reviewing the photographs. And my birthday was the 7th so the date stuck. LOL

    ReplyDelete
  13. The timing of major hurricanes making U.S. landfalls is independent of the previous landfall. There's a equal probability of a major hurricane making landfall from year to year.

    Uh, no. Just because event A does not influence event B, that does not mean events A and B are not correlated. They may both be influenced by something else, say, the North Atlantic Oscillation.

    ReplyDelete
  14. Right Wing Professor...you just proved my point. The events themselves are not correlated because there probability is a function of other atmospheric/oceanic anomalies and teleconnections. I.e. like I said, the fact that we're "overdue" for a major U.S. landfall doesn't mean there's a greater probability for one this summer because the probability of a major hurricanes making landfall from year to year in the U.S. in independent of the last major hurricane landfall.

    Wxskier...increases in SST's have an affect on the formation of major hurricanes but has no effect on their path (i.e. whether or not they'll make U.S. landfall). Also, we don't have enough historical data to make the stochastic catalogs statistically significant in understanding the probability of hurricane landfall from year to year. Yes, catastrophe models used my insurers/reinsurers do increase the probability of major landfalls but that doesn't mean they're right. I mean, AccuWeather puts out a 15 day deterministic forecast but that doesn't mean it's correct, in fact in virtually never verifies better than simple climatology.

    Mark...I'm not throwing anything at anyone, I fellow meteo alum/friend of mine posted this on facebook. I was just pointing out that it's silly to look at the graphs and make any conclusions.

    ReplyDelete
  15. Brian _

    I was just pointing out that it's silly to look at the graphs and make any conclusions

    I respectfully disagree. It might be silly to make a particular conclusion but may well be sensible to make another. The silliness is in the potentially faulty relationship between the graph and a conclusion.

    As usual "It depends".

    P.S Congratulations Dr Dr :)

    ReplyDelete
  16. Poster Brian. Exactly.. Climatology is what should be used, but the risk modeling companies have developed catalogs that imply a new climatology of landfall based on SST only, which has not verified yet. I also disagree that we don't have enough historical landfall data to develop statistically significant catalogs from year to year. Just look at the ENSO cycle. Again my point is just because SST are warmer does not mean we will have more landfalls or stronger hurricanes making landfall. In fact science is saying the opposite that strong hurricanes will weaken before landfall due to upwelling before the storm makes landfall which has been the case in several recent hurricanes, Ike, Katrina, Rita.

    ReplyDelete
  17. Readers of this thread may be interested in this paper on landfall predictions in the 1 to 5 year year time frame, such as those offered by cat modeling firms via their shorter-term catalogs:

    http://sciencepolicy.colorado.edu/admin/publication_files/resource-2786-2009.47.pdf

    ReplyDelete
  18. Right Wing Professor...you just proved my point. The events themselves are not correlated because there probability is a function of other atmospheric/oceanic anomalies and teleconnections. I.e. like I said, the fact that we're "overdue" for a major U.S. landfall doesn't mean there's a greater probability for one this summer because the probability of a major hurricanes making landfall from year to year in the U.S. in independent of the last major hurricane landfall.

    Brian: this is a respectful blog, and therefore I will only say the above betrays a basic misunderstanding of statistics and probability.

    Let's say the probability of it raining at your locality at any given instant is 5%. Let's now say it was raining five minutes ago, four minutes ago, three minutes ago, two minutes ago, and one minute ago. Is the chance it is raining right now still 5%?

    ReplyDelete
  19. Bad example but I understand your point. My point is that just because it's been a long time since the last major landfall our chances for one this summer haven't increased. The probabilities are virtually the same from year to year statistically speaking. When teleconnections are taken into consideration those probabilities do change but in general there are no statistical correlations between previous landfalls and timings of future ones.

    ReplyDelete
  20. That's a matter of observation and analysis, and can't be assumed. It may well be true, but if it's true, one needs to demonstrate it, not simply assert it.

    I did some quick statistics on landfalls of major hurricanes since 1960, and I admit it's not far from a Poisson distribution and there's no persuasive evidence of correlation (although 2004-2005 is a definite outlier). On the other hand, there is published work showing correlation with multi-year oscillations, which would require year-to-year correlation to some extent at least.

    ReplyDelete