Climate Audit

11 Comments

1. Steve Sadlov
   Posted Jan 11, 2007 at 10:08 AM | Permalink

   1851- 1900: Observations limited to widely scattered ships and shore observations. Instrumentation has large measurement errors and calibration is often questionable.
   1900 – 1950: Increasing number of ships and the advent of aircraft observations, but aircraft have limited range and cieling. Shore observations start to become more technologically sophisticated. At end of period, the first computers appear. Instrumentation has acceptable errors and calibration is required.
   1950 – 2000: Advent of satellite observations, aircraft have sufficient range to cover Most of Sargasso Sea, all of Gulf of Mexico and all of Caribbean, cielings high enough to make reasonable upper troposhere measurements on site. Automated shore measurements, digitized data, essentially unlimited data storage, computers and advanced algorithms and filters. Advanced gage R & R techniques understood (but underutilized).

2. Steve Sadlov
   Posted Jan 11, 2007 at 10:11 AM | Permalink

   Vis a vis non shore based measurements, most measurements during all periods are made by US and Canadian based groups. Other countries in the Americas play a very limited role.

3. jae
   Posted Jan 11, 2007 at 10:40 AM | Permalink

   I would say that the undercounts in previous periods are much more than one or two storms per year.

4. Brooks Hurd
   Posted Jan 11, 2007 at 11:25 AM | Permalink

   This graphic displays the impact of the change in methodology over 150 years, not only in the number of TCs recorded, but also in the vast improvement in our ability to capture TC tracks.

   Thanks, Bob K!

   I am confident that archived data of Pacific and Indian Ocean TCs would show a similar methodologcial improvement.

5. David Smith
   Posted Jan 11, 2007 at 2:48 PM | Permalink

   Prior to recon flights, circa 1950, there was little way to tell the strength of an at-sea hurricane other than from a ship’s chance encounter with the eye of the storm. Those were rare. So, just how were those at-sea red dots determined for the two pre-1950 graphs?

   Answer – they guessed the windspeed. And guessed and guessed and guessed and guessed and guessed some more and then guessed again.

6. Bob K
   Posted Jan 11, 2007 at 3:24 PM | Permalink

   Wow! I didn’t expect front page exposure. LOL

   Here are some more animated gifs of the other classes. They clearly demonstrate the differences in detection over the last 150 years. Green for TD/TS and Blue for Cat 1,2.

   Cat 1,2 anim. http://img165.imageshack.us/img165/21/cat12animug4.gif
   T Storms anim. http://img165.imageshack.us/img165/4167/tstormsanimmb3.gif
   T Depressions anim. http://img165.imageshack.us/img165/3625/tdepressionsanimjc1.gif

   Take note of the total plots for Cat 1,2 and compare to Cat 3,4,5. In the 1851-1900 period Cat 1, 2 plots are higher than the other two periods. Yet is way down in Cat 3,4,5 plots compared with the other periods. The only reasonable explanation I can find for this is lack of ship/shore communication. Ships caught in 3,4,5 may not have had a very high survival rate. Also, what captain isn’t going to steer away from the storm when encountering a severe one.

   With 1851-1900 likely severely under-reported, I tend to think those 50 years were likely the worst of the last 150. Can’t prove it though.

   The TD/TS coverage has increased dramatically as the years have passed.

7. Willis Eschenbach
   Posted Jan 11, 2007 at 5:30 PM | Permalink

   Bob K, that’s stupendous. Should be required viewing for all hurricane scientists.

   w.

8. jae
   Posted Jan 11, 2007 at 5:48 PM | Permalink

   I’ll chime in, too Bob K. That’s one of the neatest ways I’ve ever seen to display data differences over time. Wow and thanks!

9. Bob K
   Posted Jan 11, 2007 at 6:07 PM | Permalink

   Thanks fellas. I guess my brain hasn’t totally ossified yet.;)

   I just extracted initial storm readings and final readings. So here are some figures to contemplate. With so many storms in the 1st 50 years not being detected until reaching hurricane strength, it seems more than likely that many storms went unrecorded.

   1851 to 1900 storms = 377
   30=8, 35=102, 40=81, 50=63, 55=1, 60=22, 70=54, 80=14, 85=1, 90=28, 100=1, 110=2
   Mean initial speed = 51
   most common = 35
   ————
   20=2, 25=19, 30=42, 35=48, 40=69, 45=7, 50=61, 55=1, 60=42, 65=4, 70=55, 80=12, 90=14, 110=1
   Mean final speed = 49
   most common = 40
   ___________
   1901 to 1950 storms = 391
   20=1, 25=3, 30=41, 35=233, 40=37, 45=8, 50=20, 55=4, 60=39, 65=5
   Mean initial speed = 39
   most common = 35
   ———-
   15=24, 20=40, 25=82, 30=109, 35=58, 40=23, 45=16, 50=12, 55=7, 60=8, 65=4, 70=4, 75=1, 80=2, 85=1
   Mean final speed = 32
   most common = 30
   ___________
   1951 to 2000 storms = 499
   10=2, 15=6, 20=30, 25=255, 30=151, 35=29, 40=7, 45=6, 50=10, 60=3
   Mean initial speed = 28
   most common = 25
   ———–
   10=5, 15=18, 20=53, 25=111, 27=1, 30=81, 35=59, 40=38, 45=47, 50=25, 55=14, 60=26, 65=10, 70=7, 75=2, 80=2
   Mean final speed = 35
   most common 25

10. David Smith
    Posted Jan 11, 2007 at 7:22 PM | Permalink

    Great plots, Bob K!

    There’s an old American TV sitcom named “I Love Lucy”. Sometimes Lucy had some ” ‘splainin to do”. I think that anyone choosing to use the historical storm count has some ‘splainin to do about the pattern changes.

    And if they do, they need to wary of Occam and his razor.

11. Brooks Hurd
    Posted Jan 11, 2007 at 7:23 PM | Permalink

    Bob K, those are great graphics.

    I wonder what Kerry Emanuel will have to say about these graphs.

    I also wonder how we can explain the significance of these graphs to Al Gore. His film made a clear point of how we have had a large increase in TCs recently.