Kelton Halbert presents "Simulation and Analysis of Severe Thunderstorms On Blue Waters"
Вставка
- Опубліковано 6 вер 2024
- Kelton Halbert, UW-Madison graduate student working with Leigh Orf, presents at the (last - sniff sniff) Blue Waters Symposium in Sunriver, OR on June 4, 2019.
Do you have some sort of proprietary visualization code? If so, would love to hear a talk from the people responsible for it.
All of the visualizations you see from me - with only a tiny exception - are made from open source code (VisIt from Lawrence Livermore, VAPOR from NCAR, and Paraview from Kitware). The key is knowing how to use the software, basically, in an effective way that conveys what you believe is most important. Any visualization however involves a lot of choices that produce a display that contains subjective elements. Another way of looking at it is if you are sneaky and clever you can use visualizations to mask things that you know are icky in your data, or to provide a misleading interpretation. I have no interest in producing anything proprietary and share all of my code and am happy to share my methods. In fact I am going to get a video capture box soon that can just save my desktop screen to another computer while I use it so I can demonstrate the more interactive software. Stay tuned...
@@LeighOrfsThunderstormResearch This is great-thanks so much.
Pecos Hank brought me here, and I have not been disappointed. Excellent and exciting content
Me too!
@@ElectricUniverseEyes he's the best
Excellent Kelton!
Looking forward to future models!! thanks for your research.
Leigh, if you have never listen to it, I highly recommend Richard Feynman's talk _Los Alamos From Below,_ in which he describes his experience at Los Alamos during the development of the bomb. Aside from being highly entertaining, as Richard always was, I think you'll be absolutely fascinated at his stories about the trials and tribulations, and the clever workarounds, while the group worked through the modeling calculations. It will provide a richer historical context to your own work.
ua-cam.com/video/uY-u1qyRM5w/v-deo.html
Thank you for the recommendation, Feynman seems to have been a pretty cool dude as well as being a freaking genius.
@@LeighOrfsThunderstormResearch Amen to that! Plus this working all the work you're doing is just incredible oh, you are absolutely on the right track and that's probably, well I know it is, a huge understatement! When I was a kid when I was first learning about tornadoes, it never felt like the information was right and that there had to have been multiple structures combining together. Then watching the Moore tornado in 99 and seeing the first horizontal vortices live on TV. That told me a ton right there and recently what I heard about the cool pool being more important than the classic warmer and RFD type stuff, it just adds to my initial feeling of that has to be right. Actually there is no gut feelings in science other than how you should apply a method but I digress :-) amazing amazing amazing work especially the stuff you've been doing with Hank verifying in the field , this is going to revolutionize the field and I believe it already is starting to!
really interesting and thank you so much for sharing this amazing knowledge!
Back in December 10th there was the Western Kentucky Tornado that tracked for 165miles, was 1.5 miles(2,600yards) wide had winds of 190mph in Mayfield and Bremen Kentucky. This Tornado also Features the Cylcoidal marks on the ground majority of its life. The violent Tornado previous to the Kentucky one was multi vortex and had subnotical vortices. There's only one or 2 footages of the Kentucky one because of the big time storm chasers gave up on the cell since it was night and moving too fast. The reason to believe it was a multi vortex that could have been EF5 but didn't get the rating due to the poor quality construction. This simulation demonstrates the sheer intensity of Long Track long lived violent Tornadoes that made history.
11:54 I'm probably not the first person to notice this, but looks like the rain-cooled downdraft air spawns a series of mini cold fronts that are plowing into warmer air inducing uplift, and that the tornado occurs right around where it looks like the occlusion's triple point is. I'd guess that the tornado dies off when it drifts away from the triple point, just a mesocyclone starts to die shortly after the center of low pressure drifts away from the system's triple point.
Leigh Orfs Thunderstorm Research i have an idea on how the SVC forms. These misocyclones that initiate on the inflow notch or tail cloud probably lower the air pressure slightly. Those two seperate airmasses want to cancel out the slightly lower pressure, so they fill with an very slight upward movement the air. They will start mixing together and because the denser air falls down quicker, wind shear is created. And this wind shear will stack up in an long horizontal tube, which gets ingested by the inflow. Those misocyclones will get bend sideways by the horizontal vorticity tube (The SVC), helping to make the air pressure even lower than the SVC does on its own. Notice as the tornado gets stronger, these misocyclones get weaker, so its no surprise that the air pressure in the SVC gets slightly higher than when the tornado was weaker. So the only relevant information is the strenght of ambient misocyclones and the streching term along the inflow tail.
thunder heads are the result of thermocoupling between the cold dry air and the warm moist air at the boundary layer between the air masses , as the front moves the two dissimilar conductive zones get shoved across the earth's electrical field resulting in thermocoupling all the rest flows from that.
At around the 16:50 mark the subject of using computer learning to identify features came up, but as one who has been peripherally involved in such projects as classifying galaxies or mapping craters, I can tell you that at this point in time, the rule of thumb is that large numbers of untrained/minimally trained humans are MUCH better at that sort of thing than computers are, at least initially. That said, at the end of the day these are numerical simulations which may have a far more easily recognizable format that can be selected for since they are after all made of numbers & sets or ranges of numbers that can be selected for (perhaps), unlike trying to identify rocks & boulders on a newly visited asteroid or whether a galaxy is a spiral or barred spiral, for instance.
You might consider contacting Dr. Pamela Gay & her organization at CosmoQuest, as well as the folks at Zooniverse (which she was also involved with as it got up & running some years ago). They may have some thoughts which turn out to be useful once it does come time to start training AI to classify your data sets. Then again, what they do might have no bearing at all, I just thought I'd toss it out there as a lead to follow, if needed/desired. Keep up the good work!
We are definitely interested in AI to do useful things. I (virtually) attended a recent NOAA AI workshop www.star.nesdis.noaa.gov/star/meeting_2019AIWorkshop.php and it is definitely early days, but we are paying close attention to work in this area.
Wow! That sounds like an awesome job
Leigh Orf's Thunderstorm Research You are the best!
I would like to know if the rate of condensation is closely related to the drop in pressure in the tornado. If so it would help explain how so much energy can be released in a small area compared to the energy density of the whole storm. After all, evaporation stores energy and causes high pressure while condensation releases energy and causes a drop in pressure. I also wonder if dust particles sucked up by the tornado help increase the condensation rate.
Damn I wish I had that type of power.
200m sims go wheee
Very intriguing.
I would actually like to see some of the non producing simulations to compare producing ones too.
Suggest you see Brad Madsen's You Tube video "Tornado Near Miss". At around 2:15 begins what appears to be a manifestation of the SVC.
Incredible graphics...than you Leigh, I suggest tornados are ElectroMagnetic Vortices aka Lorrentz Force planetary electricity! Walter Russell
Farticus Magilicuty I concur!
Not sure if you read the comments dr leigh but... Just want to ask this but will you eventually try to simulate with the correct topography? i herd it is very important with why there are some hot spots and why places like moore keep on getting hit , with simulating just a flat ground might cause a tiny little incorrect simulation compare to the real life storms.
Hey Josh, I read 'em but sometimes it takes a while to respond. Yes, we understand the effect of not including topography. It's a difficult problem. However, the topography of places like central OK is pretty flat, so assuming a flat ground isn't probably too bad of an assumption. However, in the southeast, it's less accurate to have a flat bottom boundary. The effect of topography has been studied a bit in idealized simulations, and as computers get even more powerful it will become possible to include things like buildings and trees and hills and valleys. We're just not there yet, though, but that's the way the arrow is pointing!
few more questions but ...
1:i herd you tried the pilger event and 2013 el reno event but it didn't work out good yet , if im correct both the supercells events were attached or absorbed other storms during the tornadoes , i haven't seen you try to make multiple supercells , im wondering how other supercell/storms interacting with the main one will do in the simulation.
2:the way i look at all the small vorticies im having a hard time imagining how a gustnado would be in this simulation... have you seen any gustnadoes in your simulations? or are they secretly just tornadoes with very weak rotation at the bottom of the clouds?
3:since pilger and el reno were tried are you gonna try the dodge city 2016 event , Elie Canada F5 , the Kaite - Sulfer family along with the strange long live anticyclonic tornado in the forward flank downdraft and the jarrell tornado? the events i listed did a bunch of strange and odd things (along with the pilger/el reno 2013)
4: eventually are you gonna try to recreate a Tornado so perfect it forms, moves, looks, and ends in the same way , time and area? , i feel like the Moore 2013 EF5 is the best one to try this one out.
5:As for the SVC are drones the best way to study that spot?, i seen one guy try it out, if not what do you think is the best way to study the SVC
6:Also there was 2 possible events of a roll cloud tilting and becoming a tornado (one unconfirmed since it did it on the water) and the other was Rated EF0 happen on May 25 2008 in Bison, Kansas , is this related with the SVC in any way?
@@mario927265 We get rotation at many scales. We get 'gustnado' strength vortices - they are all over the place. They often do not cause damaging winds however. We will try more high profile events, yes, it's just a matter of time. It is one goal of mine to bring many historic (and future) storms to life and compare to observations. Drones ore maybe UAV's are probably the best way to study the SVC - see the TORUS field project information. Dunno about your roll cloud situation. Each storm is its own unique thing. We are first just trying to understand the one (well, two) we got.
i think it was one of Skip Talbot's videos.... where he talks about a tour group driving into a tornado
.
but the first tornado dies out..... and the 2nd one forms in a "coolie"
like... right in the center of the "low spot" in the coolie.... BANG... tornado forms right their
.
my guess is that the coolie focused the winds
.
ive been driving and the wind is pushing me around a bit..... but then i get to where the road on on top of a coolie
and MAN!!!! that wind REALLY tries to blow ya off the road
.
so like... while wind speed may be an average of...say.... 50mph
in the coolie... it might be 65mph
.
and if the coolie is "angled" right.... like in Skip's video
it just might help "spin up" a noodle
.
.
but im an auto mechanic.... who is just fascinated by weather
so take my "theory" with a VERY small grain of salt
Will the resolution make a significant difference when trying to simulate tornados like the 31st May 2013 El Reno tornado or other huge multi-vortex tornados?
We won't know until we try. So far I haven't gotten a satisfactory result at 30 meters. 31 May 2013 was a very complex event. It may require a more complex initialization than I usually use, and a better set of initial conditions. It was a really super rare kind of storm so it's going to be super hard to simulate "correctly"
Huh, same first name and same initials, both Meteorology related majors. Small world.
Audio sucks
Video blows
@@LeighOrfsThunderstormResearch Tornadoes do both [citation needed].