The theory of Prof. Aguste Meessen published well before Wu's article in Nature already explains ball lightning as a plasma bubble [1]. It is surprizing that Wu didn't reference this work in its introductory review.<p>In its most simple form the thin layer of the bubble is a plasma made of electrons oscillating radially. Like soap bubbles, a plasma bubble may have multiple ecapsulated bubbles.<p>A plasma ball is a dynamic system at an equilibrum using ions like ozone as energy source. The plasma ball will follow ions gradient. It explains why lightning balls may follow complex path like circling in a room and suddently going through a cheminey or a key hole.<p>He explains why plasma balls can traverse windows which are electrical and chemical reaction insulators.<p>This theory is now ready to be tested experimentally.<p>[1] <a href="http://www.meessen.net/AMeessen/BL-Theory.pdf" rel="nofollow">http://www.meessen.net/AMeessen/BL-Theory.pdf</a>
I witnessed what I think was ball lightning in my front yard as a kid, here in South Florida where lightning storms are a +thrice weekly event.<p>What happened was that lightning struck the ground at the edge of our yard while we were waiting at the front door for the rain to slow (so we could get to the car). I jerked my head to the right because of course the sound and flash scared the heck out of me, and right as I looked there was what looked like a glowing ball of electricity, the size of a beach ball, just above the ground, for what seemed like a few seconds.
This is the reference 6 of the spectrum recording of ball lightning from 2012: <a href="http://physics.aps.org/articles/v7/5" rel="nofollow">http://physics.aps.org/articles/v7/5</a><p>Impressive that the model has "passed all the tests" for an acceptable theory of the origin of ball lightning.
nit: would be helpful if the abstract or introduction gave numbers on how frequently ball lightning occurs, instead of just saying "rare". Is it 1 in 1,000 rare or 1 in a billion rare?<p>From some Googling, there are 8M lightning strikes per day on earth, so say about 3.5B strikes per year. Let's say 3% of those strikes are in populated areas. So ~250,000 lightning strikes per day in populated areas. In the past 5 years, that is ~500M lightning strikes in populated areas. Let's say that there's at least one person or video camera filming the sky during 20% of those strikes.<p>Given we only have 1 video of marginal quality of ball lightning, that makes 1 in a 100M strikes a reasonable guess. But that also makes me think it really could just be 0 and eye witness accounts are just some visual misperception.
tl;dr: Under the right conditions the step leader of a lightning strike can have a small bunch of electrons out in front of the rest of the tip, and the extremely strong electrostatic forces at play can then accelerate that bunch ahead of the tip to relativistic velocities. The bunch then hits the ground and releases an intensely strong (hundreds of gigawatts) coherent microwave radiation pulse. Also under the right conditions this microwave pulse can get trapped as a standing wave inside of a plasma bubble which it maintains through its own energy, lasting for several seconds. Scientists should be able to test this theory and create ball lightning artificially if they can build microwave devices in the 100 GW range.