The videos below were all included in my Ph.D dissertation.  That (really long) document provides a bunch more information on what's going on in each one.  Here, I'll just give a quick overview of what it is you can see in the videos.  Mostly, the videos are just cool to watch.

2012/06/22 Intra-cloud

This was the very first lightning flash that Manabu and I were able to record using the then brand new broadband lightning interferometer.  As a result, it holds a special place for me.  It also gives me a fuzzy feeling to look back on this flash, and see just how far we've come in improving the interferometer maps.  The format of the video for this flash is a little different than the rest.  The frame rate varies, so that fast moving processing can be seen without the video having a duration of several minutes.

This wasn't the first storm season we operated the interferometer, but it was the first season in which the system worked.  There were a lot of bugs in this early system.  The false trigger rate was really high, and every time the system triggered, there was about 30 seconds of dead time.  In all, we recorded 3 flashes of note that season.  But all of them were interesting.  This flash is the topic of Stock et. al, 2014, and is also discussed in Akita et. al. 2014.  It was also the flash I used to tune the early version of the processing algorithm which turns the radio signals into maps of lightning.

This flash is a good example of a fairly normal, everyday intra-cloud lightning flash (a flash which stays inside the cloud).  If we had photographed this flash, all we would have seen was a diffuse glow in the cloud.  One of the major benefits of using the VHF to map the lightning is that we can see through the cloud.  By good fortune, the flash is oriented almost face on to the interferometer.  When lightning leaders move towards or away from the interferometer site, there's a lot of perspective distortion in the maps.  In this case, there is almost none.  

The flash initiates with an upward propagating leader which carries negative charge (a negative leader).  This leader is propagating into a region of the thunder cloud which carries positive charge, since opposites attract.  After a short delay, there are several branched channels carrying positive charge which propagate downward into a region of the thunderstorm carrying negative charge.  As the channels carrying positive charge get longer, they destabilize and stop being conductive.  When this happens, a fast breakdown process travels back along the channel, warming up the air and making it conductive again.  

2012/07/23 Intra-cloud

In most ways, this flash is very similar to the one recorded on 2012/06/22 (above).  It is a 'normal' intra-cloud flash, the most common type of lightning flash there is.

The flash starts with a negative leader going up, followed shortly after by a positive leader going down.  This time, the structure of the flash is very simple, with only one primary channel going up, and one going down.  On the downside, the flash has a pretty large non-perpendicular component to it's development.  As a result there is some perspective distortion.  

in 2012, the flat plate antennas were located about 8 inches off the ground.  While that doesn't seem like a lot, it's enough that there is a blind spot overhead.  When a radio source is in the blind spot, the antennas receive no signals.  In this flash, the negative leader makes a sharp turn towards the interferometer, and in doing so, enters the null.  When this happens, the leader almost disapears in the interferometer map.  In 2013, the antenna design was changed to reduce the size of this null to almost nothing.  

2013/07/08 Hybrid Cloud-to-Ground

In Flash C, the format of the video has changed again.  Flashes A and B showed azimuth vs. elevation, in Flash C the map is shown in the direction cosine projection.  In this projection, the solid outer circle is the horizon, and the center of the circle is the zenith.  The image has some radial distortion (r is not linearly proportional to elevation), so the dotted lines show 30* and 60* elevation rings.  One of the reasons to use the direction cosine projection is that the location error in the map is the same everywhere, which is pretty cool.

Flash C connects to the ground, so it is a cloud-to-ground flash.  You can see the connection point in the upper right hand corner.  But, Flash C begins its life as an intra-cloud flash (you can tell by how the initial leader acts, but we also have 3D maps for this flash to confirm things), which is why it is labeled "Hybrid Cloud-to-Ground".  Flashes like this are also called "Bolts from the Blue" because the channel to ground can travel a long way horizontally before touching the ground, far enough that it sometimes comes to ground outside the flash.  

When the channel touches ground, an EM wave (called the return stroke) travels up the channel into the cloud, bringing ground voltage potential with it.  Because the potential difference between the cloud and the ground is around 100,000,000 V (a lot), when the return stroke gets to the end of the channel there's a bright (in the VHF at least) burst of activity.  This burst is seen in most all cloud-to-ground return strokes, but tends to be especially impressive in hybrid bolt from the blue type flashes.  

2013/07/24 Low Altitude Intra-cloud

Classic thunderstorms have 3 primary charge layers, arranged kinda like pancakes.  Usually, the lowest layer is positively charged, the middle layer is negative, and the upper layer is positive again.  Usually intra-cloud flashes initiate between the upper two layers, and the preliminary leader carries negative charge going upwards (that's Flash A and B).  Because the lower positive charge layer is usually pretty small, when a flash initiates between the lower two layers, the channels frequently continue on to ground, resulting in a cloud-to-ground flash.  But, sometimes the lower positive charge layer is larger, and the flash doesn't continue to ground.  This is what happened in Flash D.

There's a big difference between how the channels extend in low altitude intra-cloud flashes and high altitude (normal) intra-cloud flashes.  The channels carrying negative charge seem to propagate faster, step more frequently, and branch a lot more in low altitude flashes.  Because there are more channels, frequently more than 1 channel is producing VHF radiation at any given time.  When these very few element interferometers, this causes some smearing to happen in the maps.  

I'm not sure exactly what causes the difference in behavior, but a good guess is that it has to do with the pressure.  Flashes happening at lower altitude happen at higher pressure, where the mean free path of an electron is shorter.  But I don't think anyone has confirmed that's what's going on.