Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Tornados Starting from the Bottom-Up?

  1. Dec 15, 2018 #1


    User Avatar
    Science Advisor
    2018 Award

    Not how I have always heard these things get started from a high level:
    Science News article
    Cases were documented where wind spinning starts first near the ground.
    They were able to see this using a fancy mobile radar from a high vantage point.
    This allowed them to see lower altitudes, with the radar, than 250 meters.
  2. jcsd
  3. Dec 15, 2018 #2


    Staff: Mentor

    Yes, this was quite a shock. Lightning works in a similar fashion and was shown to do this via some super high speed camera work a few years ago.
  4. Dec 16, 2018 #3


    User Avatar
    Gold Member

    Is the implication that this is the typical way tornadoes form, or is this considered the exception?
  5. Dec 21, 2018 #4
    Most tornadoes form high in the sky over land and are top-down.

    Most waterspouts form low in the sky over water and are bottom-up.

    There are many exceptions to these two rules of thumb.

    "Gustnadoes" are essentially large whirlwinds that become powerful enough to be classified as low-strength F0 or F1 tornadoes.


    "Landspouts" are weak tornadic storms that form without a mesocyclone over land.


    Likewise, it's common to see mesocyclonic tornadoes form over water such as the Great Lakes or oceans when thunderstorms roll across them, and the mechanisms are mostly the same. A thunderstorm that is already 30,000-80,000 feet high that rolls across a lake or river will not change its form much simply because the surface at its base is water instead of land.
    Last edited: Dec 21, 2018
  6. Dec 21, 2018 #5


    Staff: Mentor

    However, the recent research article suggests otherwise that they see rotation on the ground first before they see it higher up.
  7. Dec 22, 2018 #6


    User Avatar
    Science Advisor

    Any numbers (counts) for the 'most' and 'recent' qualifiers?
  8. Dec 22, 2018 #7


    Staff: Mentor

    The article didn’t posit any counts beyond 4 instances that were observed. I imagine that more research will continue to explain the observations better.
  9. Dec 22, 2018 #8


    User Avatar
    Gold Member

    Forty years ago I worked in an open-air laboratory spread over the Sierra Nevada mountain range, mainly desert terrain. Tornadoes, clear air turbulence and wind gusts were so formidable, I wrote a technical order (TO) that unpowered radar antennas should be secured after operations, if only to avoid wear on bearings and structure.

    Operators on systems that included optical tracking claimed they watched 'dust devils' form from the desert floor. This report could verify their reports. Note: the radars were not designed to track atmospheric conditions but moving target indicators (MTI) could build interesting returns depending on conditions.
  10. Dec 25, 2018 #9
    I don't have numbers available.

    I would imagine that the easiest way to discern between top-down and bottom-up tornadoes would be to use Doppler radar to determine if, and when, and at what height, a significant amount of rotation was observed for every tornado warning issued, and then to match that data set against the list of verified tornadoes from the National Weather Service.

    If most of the tornadoes that were verified were detected by radar with significant rotation before the time that they were seen on the ground, then that would answer your question.

    That would require a significant exercise in data mining, but it would actually be a good idea for a master's or Ph. D. thesis.

    There are some areas of the US that are not well covered by Doppler radar, but they get few tornadoes every year.
  11. Dec 25, 2018 #10


    User Avatar
    Science Advisor
    2018 Award

    The article I linked in post #1 to said that normally the radar that is normally used does not get good info from below 250 meters.
    They could see the lower rotation only from a higher vantage point.
  12. Dec 26, 2018 #11


    User Avatar
    Gold Member

    That's a quote from the wiki on landspouts.
    If the bottom up is not filled with condensation, Doppler won't pick it up and the swirl will be transparent, and only visible if debris is uplifted.
  13. Dec 30, 2018 #12
    If there is no significant rotation observed from Doppler radar before a tornado, and the tornado is verified by ground observation of a path of damage, that would most likely be a bottom-up tornado. Some tornadoes might be too distant from a radar station to tell the difference, however.

    Most tornadoes appear to be preceded by significant high-level rotation, but I don't have any firm numbers to share.

    It looks like it would be easy to data mine the Doppler radar records to tell if a tornado had a lot of high-level rotation before it was spotted on the ground, but harder to discern low-level rotation just before a bottom-up tornado is created.
  14. Jan 4, 2019 #13


    User Avatar
    Homework Helper

    I just looked at several videos of tornadoes forming on youtube, and they all give the appearance of forming from the top. So there does not seem to be any issue about "spotting" them on the ground. If there is rotation at or near ground level, it is without visual affects.

    One thing that I did not see in any of the videos I was looking at was direct evidence of the ground-level winds - such as flying/breaking objects. This evidence was not apparent either as the visible portion of the tornado cloud was descending, nor was any apparent soon after it had touched down. There were cases where dust clouds formed at ground level soon after touchdown - but those dust clouds are hard to see in the videos. They would have been easy to miss as the cloud descended. Part of the problem is that before the tornado touched down, most of the videographers point their cameras high.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?