Made a simple formula

[Sam Kennedy]

I made a simple formula for finding the percentage chance of tornadoes.

It's FAR from perfect, and I've only done high school maths so I'm no pro, so I decided to post it here so people can do what they like with it.

Chance(%) = dewpoint(ºF) + cape(j/kg) + low level jet(850mb,kts) + upper level jet(200mb, kts) + vort max + helicity(0-3km) / 3285 * ((li*-1)/8) * 100

So there it is! You can do what you like with it, but if you earn any money with it I demand 55% of it!

If you need anything explained just ask

 

[Tim Vasquez]

I bet it works great but the FAR ratio might be an issue here unless the CAPE stays under 100.

Tim

 

[Chris Frost]

Hmm...I think I did it wrong. I must have forgot to carry the 1 because I just used this formula for May 29 2004 and came up with a 1200% chance of tornadoes

Although that was a heck of a day

 

[Skip Talbot]

What's vort max? Is this vorticity at a particular level? What level? What units is it represented in?

Why 0-3km SRH instead of 0-1km or effective SRH?

Can you plug in some numbers for an example?

 

[Tim Vasquez]

Here's a variation of your idea that might keep things closer to 0-100%.
% = (Td-50) + (CAPE/100) + ((V200-V850)/2) + (η12) + (SRH / 10)

(η12 = absolute vorticity change over 12 hours in regular units of 10-5 s, SRH is SRH in 0-1 km layer, V is speed in kt, Td in Fahrenheit)

There are five terms, and you can easily play around with it to change the contribution of each term by changing the coefficient of each part. The -50 in Td-50 normalizes the entire term to zero if the temperature is 50F, and anything above contributes and anything below works against the %. You can put a coefficient on that part by just writing (Td-50)/2 or some such. Adding 200 and 850 winds might be good for forecasting downbursts but here taking a difference would be better for tornado forecasting.

Anyway, there's some ideas for those who like inventing indexes. Using one term for each contribution and using normalization & coefficient on each term to change the contribution of each portion is a pretty good way to do it.

EDIT: Oops, I saw you had LI -- I forgot to add that.

Tim

 

[Keith Mullen]

ffzzzt!

Um... whoah.

I think my cerebellum just fused.

Holy moly.

 

[Darren Addy]

Humans tend to want to distill things down to a magic number but reality isn't quite that simple. I'm curious as to how some of the "established" meteorological indices were developed. Are the numbers massaged until the results appear to line up with observed realities? Some components are (like Einstein's Cosmological Constant) almost completely arbitrary.

Related to this: I'm not sure how many people have read Doswell & Schultz's paper entitled On the use of Indices and Parameters in Forecasting Severe Storms but it has some thought-provoking stuff in it regarding the pursuit of "magic numbers" and thresholds in forecasting. (IMHO the paper has some problems, like the conclusion not appearing to be in sync with the purposes stated in the introduction, but nevertheless it is a valuable read. The comments and ensuing discussion by peer reviewers at the end is especially entertaining).

An earlier paper by Doswell on this subject: On Convective Indices and Sounding Classification is somewhat related and show that Doswell's suspicion of these things has not changed over the years. (I think this paper was originally published in 1996, but the link contains a reference in the annotated bib. to a 1997 paper, so perhaps the link is to a revised version?)