• Welcome to TalkWeather!
    We see you lurking around TalkWeather! Take the extra step and join us today to view attachments, see less ads and maybe even join the discussion.
    CLICK TO JOIN TALKWEATHER

Svr Wx threat April 1-April 2 2024

NAM 3km helicity tracks for tomorrow. Considering how insipid the NAM can be when it comes to convection, the cluster of streaks over AL/GA from forecasted QLCS-embedded supercells is cause for concern. Looks like a potent situation for anyone south of the Great Lakes, but we'll have to see what becomes of crapvection in the OH Valley tomorrow.
uh03_max.us_ov (1).png
 
Look at those analog matches! Middle TN. Nasty sounding.
 

Attachments

  • IMG_3843.jpeg
    IMG_3843.jpeg
    643.2 KB · Views: 0
I also thought I'd talk for a moment about the STP values, because that seems to be a distinct tangent here. In the seminal paper that literally defined the SPC's modern definition of STP (https://www.spc.noaa.gov/publications/thompson/stp_scp.pdf), the math and methodology are discussed. In the paper, box plots of STP's are even shown --- and values of 5,6, and up are exceedingly rare. An STP of 85 (for instance) is simply nonsensical and carries no information that can translate into actionable intel. One thing to understand about the STP is that it is a *conditional* probabilistic index. The STP is NOT a direct measure of how strong tornadoes will be (if that were the case, STP's like we see from the sites posted of 12-18+ STP, we'd be dealing with EF7's...), and on the same token, the STP is not directly measuring the likelihood of any threshold of tornadoes. Instead, the STP says "IF we are given a right-moving supercell that has already formed, are we able to discriminate cleanly between that supercell being significantly tornadic or not?" Thus, an STP of 0.7 would not very cleanly separate between NON, TOR, and SIG. But an STP of 1.8 does that better. And an STP of 3.5 does that even better. But at some reasonable point around STP of 4 or 5, the increasing of STP is essentially meaningless. Once STP reaches 4 or 5, it's effectively fully separated the population distributions, historically, between NON, TOR, and SIG. So any site putting out proprietary algorithms of STP's of 15 or 20 or 80 are just driving hype and clicks, in my view. An STP of 4 is functionally equivalent to an STP of 80.

*steps off soapbox*
 
wx_guy said:
I also thought I'd talk for a moment about the STP values, because that seems to be a distinct tangent here. In the seminal paper that literally defined the SPC's modern definition of STP (https://www.spc.noaa.gov/publications/thompson/stp_scp.pdf), the math and methodology are discussed. In the paper, box plots of STP's are even shown --- and values of 5,6, and up are exceedingly rare. An STP of 85 (for instance) is simply nonsensical and carries no information that can translate into actionable intel. One thing to understand about the STP is that it is a *conditional* probabilistic index. The STP is NOT a direct measure of how strong tornadoes will be (if that were the case, STP's like we see from the sites posted of 12-18+ STP, we'd be dealing with EF7's...), and on the same token, the STP is not directly measuring the likelihood of any threshold of tornadoes. Instead, the STP says "IF we are given a right-moving supercell that has already formed, are we able to discriminate cleanly between that supercell being significantly tornadic or not?" Thus, an STP of 0.7 would not very cleanly separate between NON, TOR, and SIG. But an STP of 1.8 does that better. And an STP of 3.5 does that even better. But at some reasonable point around STP of 4 or 5, the increasing of STP is essentially meaningless. Once STP reaches 4 or 5, it's effectively fully separated the population distributions, historically, between NON, TOR, and SIG. So any site putting out proprietary algorithms of STP's of 15 or 20 or 80 are just driving hype and clicks, in my view. An STP of 4 is functionally equivalent to an STP of 80.

*steps off soapbox*
Click to expand...
Very interesting. Seems like you would have a way to put a limit on how high the values can reach. Anyways, I honestly didn't know about the different algorithms.
 
Meanwhile 18z NAM CWASP with 75 to 80 contour over parts of North MS into Alabama.
 

Attachments

  • CWASPnam212F027.png
    CWASPnam212F027.png
    272.7 KB · Views: 0
JPWX said:
Meanwhile 18z NAM CWASP with 75 to 80 contour over parts of North MS into Alabama.
Click to expand...
Excuse Me Wow GIF by Mashable
 
JPWX said:
Very interesting. Seems like you would have a way to put a limit on how high the values can reach. Anyways, I honestly didn't know about the different algorithms.
Click to expand...
Yeah, I'm not entirely sure what modifications to the math some sites use, but essentially, the basic formula is several normalized values multiplied together, so technically there's no limit to how high they can go, as long as their respective distributions and spreads support the extreme ends.
 
wx_guy said:
Yeah, I'm not entirely sure what modifications to the math some sites use, but essentially, the basic formula is several normalized values multiplied together, so technically there's no limit to how high they can go, as long as their respective distributions and spreads support the extreme ends.
Click to expand...
STP = X(5) + 2 = number go big
 
You must log in or register to reply here.
Back
Top