Hacker News new | ask | show | jobs
by asynchronous13 2571 days ago
Complete tornado records in the US begin in 1950 -- how far back does it need to go for it to be worthwhile in your opinion?
4 comments

I think perhaps there’s likely thousands of tornadoes that never got recorded due to locality in low population areas.

I grew up in tornado alley and can remember many tornadoes that did little or no damage that never got a blip in the news.

Urban areas in 1950 were small compared to today - perhaps our population and technology are just exposing more than we knew before.

That is actually addressed in czr's video link above, specifically the one minute or so from here:

https://youtu.be/Z_1PiixPX3o?t=881

The key points being:

- Low intensity tornados are not included in the statistics

- rural population in the relevant areas has actually decreased, not increased, over the time period.

- it's hard to explain the variation in metrics as being due to different reporting rates (better explanation in video)

2000 years for events like this to make a “record” significant. Particular when records are of the sort like “11th consecutive day with at least 8”.

Ever listen to a baseball game with an announcer who loves stats? They can find something about nearly every game that breaks a record of some sort.

It's true, that is one of the things about lies, damned lies, and statistics - the more creative you are, the easier it is to invent new record breaking stats.

2000 years seems somewhat arbitrary - would you apply the same threshold to "500 tornadoes within a 30 day span", what about 1,000, what about 10,000?

There probably ought to a threshold where extraordinary evidence is sufficient to make extraordinary claims, otherwise if your requirement is 2000 years, I'm sorry to report that nothing could ever persuade you about climate science, or really most science.

I was thinking more about it, feeling a bit bad about my somewhat glib comment, and I came back to add, I think as long as you can come up with equally impressive sounding statistics pushing the other side of the narrative, it’s no use.

But to your point, annual statistics showing multiple sigma deviation for several years out of a 10 year period, that to me would tells us something has changed.

Then it occurred to me, I don’t know anything about number of tornadoes per year, how much variance is in the data, and what’s the trend like, but a quick Google search brought me here;

https://www.ustornadoes.com/2012/04/10/violent-f4ef-4-and-f5...

And it sure looks like the “Violent US Tornadoes by Year” has massive variance and little to no discernible trend line.

What I learned is that almost all violent tornadoes happen in April and May and some years have huge clusters and some years have very few.

This kind of phenomenon requires a very long historical series coupled with a long stretch of deviation from the norm to be able to say anything meaningful about recent data, other than you are most likely seeing new random output from the same underlying/unchanged function.

It short, I hate news stories about the weather. They need to make something inherent about the world seem new.

EDIT: Here’s another good one from the same site;

“I decided to take a look at the number of tornadoes that have been reported in the Plains between May 16-31 over the past 25 years, from 1993 to 2017. ... The most was 204 in 2004, while the least was just nine in both 2006 and 2009.”

In that ~two week period (which coincides with the current burst of activity) we saw some years 9, one year 204, in the last 25 years.

What’s so discernible about this year in that context?

[1] - https://www.ustornadoes.com/2019/05/14/how-peak-tornado-seas...

Careful, that sort of logical thinking might inspire questions about the whole tottering edifice...
To add on to zaroth's remarks, here's a list of records set or tied in the most recent (and IMO not very remarkable) 53rd Super Bowl:

https://www.si.com/nfl/2019/02/03/patriots-rams-super-bowl-2...

"Complete tornado records" don't go anywhere near as far back as 1950.

The modern NEXRAD system wasn't even deployed until the early 1990s, and has been upgraded since then. Also, we now have zillions of storm-chasers cruising the highways any time severe weather is in the offing.

https://www.weather.gov/gld/tornado-tornadographs

> In the early 1990s, a major modernization program within NWS took place, which included the deployment of Doppler radar and a more rigorous spotter training program. In addition, the 1990s saw an increasing number of storm chasers using video cameras and cell phones to document and report severe weather. These facts are likely responsible for the dramatic increase in annual tornadoes reported since 1990.

Also, as recently as 2014, we were in a three year record low period for tornadoes.

https://weather.com/safety/tornado/news/tornado-count-hits-r...

However, that wasn't worthy of sensationalistic headlines in the NYT.

Can you explain why we should place greater emphasis on an 11 day high than on a three year long low?

> However, that wasn't worthy of sensationalistic headlines in the NYT.

The headline was formed based on the quote from the academic involved:

'We are flirting in uncharted territory,' Dr. Marsh

So if the headline is sensationalized that suggests Dr. Marsh is the sensationalist.

And here in lies the whole problem with the climate change debate.

An academic who I suspect spends a lot of his/her working life studying the actual problem it immediately written of as being a sensationalist.

But this constant shoot the messenger to climate change denial will not stop the contents of the message from playing out in time.

> The headline was formed based on the quote from the academic involved:

The headline was cherrypicked from an interview.

> The headline was formed based on the quote from the academic involved:

No, it would suggest that the reporter picked the most sensationalistic quote he could find from a lengthy interview.

> But this constant shoot the messenger to climate change denial will not stop the contents of the message from playing out in time.

I notice that you did not address any of the facts linked from my post, such as the number of tornadoes being at a record low between 2012 and 2014, nor the fact that according to the National Weather Service's own site, the increase in the number of reported tornadoes is largely due to better radars and more storm-chasers looking for them.