Violent Tornado Parameter (VTP)

Randy Jennings

Supporter
Joined
May 18, 2013
Messages
782
A new paper by Nick Hampshire (NWS EWX), Richard Mosier (SPC), Ted Ryan (NWS FWD), and Dennis Cavanaugh (NWS LZK) entitled "Relationship of Low-Level Instability and Tornado Damage Rating Based on Observed Soundings" has been published in the Journal of Operational Meteorology. It is an interesting read and it offers a new violent tornado parameter (VTP) that similar to STP but includes the low-level instability fields in order to better differentiate between significant and violent tornado environments.

Full paper: http://nwafiles.nwas.org/jom/articles/2018/2018-JOM1/2018-JOM1.pdf
 
A new paper by Nick Hampshire (NWS EWX), Richard Mosier (SPC), Ted Ryan (NWS FWD), and Dennis Cavanaugh (NWS LZK) entitled "Relationship of Low-Level Instability and Tornado Damage Rating Based on Observed Soundings" has been published in the Journal of Operational Meteorology. It is an interesting read and it offers a new violent tornado parameter (VTP) that similar to STP but includes the low-level instability fields in order to better differentiate between significant and violent tornado environments.

Full paper: http://nwafiles.nwas.org/jom/articles/2018/2018-JOM1/2018-JOM1.pdf
VTP was just added to the RAOB Program's parameter listing and export options > http://www.raob.com/raob69beta.htm
 
This is a very interesting article. I am assuming the only way to accurately use VTP would be from a second sounding on potentially high risk days given the parameters would be subject to constant DL movement, mixing, outflows, etc.?
 
Good research in an attempt to further distinguish environments supportive of high-end tornadoes from other tornadoes.

While it's nice to see some encouraging quantitative results, I don't think this paper discovered any truly new gems of information that weren't already known. Ultimately, the strength and damage rating of a tornado come down to other non-meteorological and sub-mesoscale factors, such as the distribution/concentration of DIs (not all tornadoes hit stuff, and for many DIs the maximum DOD equates to wind speed below the EF5 threshold, meaning that DI cannot be used to classify a tornado as EF5. I think some DIs even max out at EF3 winds), and the small-scale nature of tornadoes themselves. Our measurement systems are not capable of adequately sampling the scale of features most responsible for distinguishing tornadoes of various strengths from each other, such as the magnitude of the horizontal vorticity generated by the RFDGF, as well as the magnitude of the actual vertical velocity near the ground.
 
Good research in an attempt to further distinguish environments supportive of high-end tornadoes from other tornadoes.

While it's nice to see some encouraging quantitative results, I don't think this paper discovered any truly new gems of information that weren't already known. Ultimately, the strength and damage rating of a tornado come down to other non-meteorological and sub-mesoscale factors, such as the distribution/concentration of DIs (not all tornadoes hit stuff, and for many DIs the maximum DOD equates to wind speed below the EF5 threshold, meaning that DI cannot be used to classify a tornado as EF5. I think some DIs even max out at EF3 winds), and the small-scale nature of tornadoes themselves. Our measurement systems are not capable of adequately sampling the scale of features most responsible for distinguishing tornadoes of various strengths from each other, such as the magnitude of the horizontal vorticity generated by the RFDGF, as well as the magnitude of the actual vertical velocity near the ground.
It will be interesting to see what comes of the experiments OU/UNL/TTU and others are trying to do. UNL is leading the charge on RFD-sampling with UAS and I would imagine, if sufficient sampling IS possible, that some interesting findings/advancements could come of it.
 
It will be interesting to see what comes of the experiments OU/UNL/TTU and others are trying to do. UNL is leading the charge on RFD-sampling with UAS and I would imagine, if sufficient sampling IS possible, that some interesting findings/advancements could come of it.

Not in real-time, though. Great for post-mortem research, and it lays a foundation for future prediction, but we're probably decades away from being able to predict anything like this ahead of time.
 
Just a quick observation:

The VTP equation factors both 0-3km MLCAPE and 0-3km lapse rates. That seems redundant, as favorable low-level thermodynamics can skew the parameter. For example, yesterday was throwing >1 values in eastern Colorado and this morning, there are values well over 1 across much of Kansas. (yes, the environment becomes more favorable later in the day, but with such a high threshold for LCLs, strong boundary layer heating is skewing the parameter)
Untitled-4.png
Think about it this way: If 0-3km lapse rates are steep, in most severe weather setups, there is going to be sizable 0-3km MLCAPE. That's just how it works... By multiplying the values together in the equation, in my opinion from a quick look, it seems potentially flawed. My advice is that caution should always be used when reviewing severe weather parameters. I use them a lot less than I did in the past, but I do understand that in the big picture, they can be helpful. Once you start tracking wildly different types of events and consider other mesoscale factors, then their utility comes more into question.

Yes, low-level instability is favored, but I'm not sure I'll find myself using this parameter much. I hope there are case studies with VTP. I'm sure it will work in some environments, but I bet there will be a lot of false flags too. Maybe the LCL threshold should be lowered from 2000m. I always thought that was kind of high, although the paper does talk about substantial low-level instability being able to overcome seemingly high LCLs.
 
As with any composite parameter - it's just a mish-mash of root fields considered to be important in forecasting whatever phenomenon. All composite parameters are pretty arbitrary because their values don't necessarily have any meaning. What does a VTP value of 3 mean? Without a sample of data on which to base the relationship between VTP and the occurrence of violent tornadoes, nothing. Yes, I can see how the thresholds were picked to try to assert that a value of VTP > 1 is a subjectively necessary threshold to get violent tornadoes, but again, there's no guarantee the statistics will bear that out. The same goes for SCP, STP, derecho composite, and etc.

None of us have seen enough cases yet with VTP displayed to have any real idea what to make of that field. It will take time to build up a sample and develop a climatology to understand the significance of any particular magnitude.

It would probably help interested readers on this subject to consider Chuck Doswell's essays on composite indices. A good introductory essay can be found here: http://www.flame.org/~cdoswell/indices/Indices_and_Parameters.html
 
In regard to the redundancy of 0-3km CAPE and 0-3km Lapse Rates, I disagree.

For one, the 0-3km CAPE is actually MLCAPE, which comes up with a mean mixing ratio and theta in the lowest 100mb and does **NOT** include the environmental lapse rate (save for in reducing positive buoyancy). Additionally, the 0-3km Lapse Rate tells you nothing about the moisture content and low level buoyancy. The lapse rate is really also a sort of regulator on vortex stretching potential. You can have strong low level buoyancy and not the best lapse rate and that could inhibit vortex stretching some.

Just throwing in: you can access VTP mesoanalysis contours on the https://satsquatch.com application if you are a subscriber ($5/mo).
 
Last edited:
I think it's important to understand the potential pros and cons of the parameter. No severe weather index is perfect. Having this discussion will hopefully encourage people to objectively assess severe weather setups and not just focus on where indices are highest.

Based on what I've seen so far, VTP has been skewed high in environments that are clearly not favorable for violent tornadoes. Environments with large 0-3km instability and steep 0-3km lapse rates are not always favorable for tornadoes, let alone violent ones.

When one understands the caveats with any severe weather parameter, they can more effectively use it.
 
I still think a good old surface chart and visible satellite can't go wrong. Parameters will max out near boundary intersections juxtaposed with high theta E. I use parameters to cross-check my old school analysis. When they diverge, I often trust the surface/vis more.

Bottom line, I agree with Quincy that some of these just double and triple down on the same number. I have seen a paper justifying VTP compared to STP, and maybe there is value forecasting for the public. For chasing, I'm almost always OFB / DL intersection.
 
Let me reiterate that I think the concept behind VTP is great, but this particular formula needs some work, in my opinion.

If the purpose of VTP was to improve STP and/or differentiate between environments supportive of weaker vs. more intense tornadoes, I'm not sure it's doing so well. I've only been watching VTP for a few events and it's already throwing out seemingly flawed data. (I still think much of this comes back to factoring both 0-3km MLCAPE and 0-3km lapse rates, which skews the parameter in cases with substantial low-level instability, but maybe there are other issues as well)

Perhaps it's a problem with the scale. Remember that 1.0 is supposed to be a benchmark threshold, suggesting that values over 1.0 are correlated with violent tornadoes, just as STP values over 1.0 are associated with strong tornadoes. Despite that, yesterday, mesoanalysis showed VTP values (3+) that were higher than STP (around 1) in northeastern Kansas. While one could argue that maybe the environment was marginally supportive for a weak tornado, there is no way it was supportive of a violent tornado. It's a big problem when VTP is higher than STP, especially in a scenario in which a violent tornado is extremely unlikely.

The paper linked does indicate that in violent tornado environments, VTP is often higher than STP. That's fine, but if it's also higher (sometimes much higher) than STP in environments that are not particularly favorable for tornadoes, then what's the point? STP was set to 1.0 as a threshold and maybe that's what should have been done with VTP.

I'm kind of scratching my head here, because the paper states:
...there were negligible differences between the STP and VTP for the weak and significant tornado database
I would also expect STP and VTP to be similar in low-end or near-zero tornado threat areas. I still argue that VTP should be lower than STP in those cases. It's like saying that the majority of violent tornadoes had a higher VTP value than STP, but in many (perhaps very many) non-tornado environments, VTP was also higher than STP.

Basically, if there's much low-level instability at all, VTP>STP. That's ripe for false positives.

I think the parameter is going to be throwing up a lot of red flags. I had someone suggest that it's only valid in right-moving supercell cases. Well, yesterday had the potential for supercells in northeastern Kansas, but maybe it wasn't the best example. I'm not sure it's worth the effort, but I kind of want to just compare VTP and STP in as many events as possible, but I'm not sure who I'd be trying to convince...

I will continue to assess the parameter as much as I can through this spring and share findings. So far, I'm not impressed.
 
I'm familiar with the early stages of this work. Upon further reflection, probably the best way to approach the VTP would be to *reduce* STP (as mentioned above) in areas where the low-level lapse rates and low-level CAPE are small. Instead of allowing values to get larger through the lapse rate and CAPE terms, you could simply set VTP=STP when the 0-3 km lapse rate >= 6.5 C/km and 0-3 km MLCAPE >= 50 J/kg.
 
Back
Top