In which I discourse about hurricanes and global warming with Sarah Goforth of Discovery News:
What do folks think of the ending, where--following my lead--the video contrasts the number of Category 5 storms between 1970 and 2002 (8) with the number between 2003 and 2007 (7)? My own view is that while we shouldn't make too much of a comparison like this, it is still worrisome...
- Log in to post comments
Marginal. When I see that sort of comparison, I always want to go a level lower to check that the selection of time periods hasn't been fudged.
Example - if there had been no Cat 5's between 1989 and 2002, then one could also have said - "7 from 1970 to 1989; 8 from 1990 to 2007 - big deal."
I find the "4 in 2005 where no other year since 1970 has had more than 1" probably has more weight with me. Even there though, I wonder, "Why the 1970 cutoff?"
I imagine the cutoffs generally involve either some form of accuracy (if wind speeds could not be properly registered < 1970), or some form of definition. I don't know enough about hurricane science to discuss why particular dates are used for statistics. All I know is 30 years isn't a whole lot.
I understand why they did it, but I prefer graphs that show as much of the available data as possible...
Regardless, you again did a good job.
My point was not to wonder what the reason was, but to illustrate why it's worthwhile to read these sort of "simple stats" with a critical eye.
In this case, what if there were 4 Cat 5 hurricanes in 1969?
I'm not suggesting that Mooney is "cooking the books," just illustrating why I don't automatically accept stats that are potentially so easily manipulated.
Scott -- You're right on the money. 1970 sounds like a nice round cut-off, doesn't it? Except that 1969 previously held the record for most hurricanes in a season (until the outlier year 2005). 1969 was also memorable for the catastrophic Hurricane Camille, a Category 5 that made nearly the same landfall as Katrina and killed 259 people.
very good progekt.