• Stormtrack's forum runs on Xenforo forum software, which will be undergoing a major update the evening of Wednesday, Feb 28th. The site may be down for a period while that update takes place.

'Anvil zits' lightning indicative of rotating updraft?

Dan Robinson

I know the CG lightning-tornado relationship theory has been rehashed frequently, but here's a little varaition on this topic. In every classic supercell I've observed to date, there has been the presence of the small, nearly continuous lightning discharges at high altitudes in the updraft tower, accompanied by soft, continuous thunder. The lightning channels are usually a deep blue or turquoise in appearance. I've heard these referred to as 'anvil zits'. Here are some example video/still frames:


I'd be interested to hear if anyone else has observed any consistent relationship between this type of lightning and supercells that produce tornadoes, or at the least with supercells with strongly rotating updrafts.

I can only think of two instances in the past five years where I heard/saw this type of lightning/thunder and did *not* see a tornado with the same storm (5/27/02 Crosbyton, TX is one example). And in *every* instance I have seen a tornado, I have observed these 'anvil zits' and their characteristic soft, continuous thunder. This was especially true this season, when I took note of every time I observed 'anvil zits', and in every instance I witnessed a tornado. There were storms that did not exhibit this type of lightning (6/13/05 in Rolla, Missouri was one example) that failed to produce. The Rolla storm did produce a considerable amount of cloud-to-ground strikes.

With Mulvane 6/12/04, of course, there was considerable CG activity in the updraft vicinity of the storm (around the tornado) but the 'anvil zits' were still there as well. Their presence on 6/12/04 (and 6/12/05 in Kent County, TX in another example) was overshadowed largely by the more 'vocal' CG barrage, but they were still there (as darkness fell, it was easier to see them up there).

These 'anvil zits' are usually, but not always, in the immediate area above the updraft base (not in the forward-flank precip or anvil regions away from the updraft). There is usually, but not always (an exception is the 6/12/04 case), a small or nonexistent amount of CG activity in the updraft region - CG activity is largely confined to the forward-flank precipitation region while the 'anvil zits' are ongoing.

In fact, the main visual association I have with hearing this soft, continuous thunder overhead is the one of condensing wall clouds, tail clouds and finally a funnel dropping.

My main question, in a nutshell, is if the presence of this type of lightning is a reliable indicator, in itself, of a rotating updraft, and specifically, one that is more likely to be tornadic.

I remember that there was a lot of discussion during May 2003, but the discussion (IIRC) was about the LACK of lightning in tornadic supercells. On 5-8-03, I was on the northern OK / Osage county tornadic supercell, and I witnessed only 1 or 2 lightning bolts the entire chase, which included an hour or so after sunset. There was simply no lightning at all with this strong tornadic supercell. I don't recall much lightning either with the 5-4-03 ne OK to SGF tornadic supercell. In contrast with 2003, it seemed that the 2004 storms were very electrically-active -- 6-12 Mulvane supercell being a very good example of this.

First of all, updraft rotation, which characterizes a supercell, enhances/strengthens the updraft to a point that rotation may contribute more to updraft strength than buoyancy (CAPE). Since lightning is based on charge seperation (ice particle seperation), a strong updraft helps accomplish this.

What was different between 2003 and 2004 May outbreaks? A lot of the 2003 outbreaks were accompanied by very strong mid-upper level flow (remember the 60-80kts at 500mb and 80-120kts at 250mb?), while many of the 2004 outreaks were more on the 40-60kts range in the mid and upper levels. Perhaps the 2003 supercells faced so strong mid/upper level SR winds that the charge seperation was too much... I duno, just a guess.

For what it's worth, the April 21st 2005 se KS / sw MO HP was the most electrically-active storm I have ever chased. The CG activity was incessant and absolutely incredible. To support my guess, the 250mb flow on SGF sounding that evening was <=30kts...
Interesting topic, Dan ... I think we see the 'lightning around the meso' topic a lot because it is something the chasers encounter frequently, so naturally there is a lot of discussion. As for the anvil flickers you have noted above the primary updraft, I have not personally seen something similar and made a mental connection with rotation necessarily (usually because if there's a rotating updraft involved I can't maintain a clear view of where the updraft intersects the anvil). I have seen similar lightning once I allow the storm to move on and I move back out from under the rear flank, however.

I have a personal theory based on observation concerning lightning interaction with rotating supercells. I do not personally believe that there is or should necessarily be a direct correlation between a rotating updraft and high lightning production in the early stages of a supercell. There are two reasons I say this:

1. Mini supercells produce little, if any lightning and yet remain able to maintain a high possibility of producing tornadoes at times. I have personally seen this on multiple occasions - sometimes there is NO lightning/thunder AT ALL, and yet here we still have a perfect mechanism for supercell-spawned tornado.

2. Tornadoes that are spawned from true supercells often occur as soon as the updraft reaches maturity, but before serious lightning production begins. Examples: 5/22/04 Hebron storm (Hallam) ... tornadogenesis first occurs as soon as the reaches a balanced maturity, but we do not start noticing lightning around the mesocyclone for maybe 20-30 minutes following the first true tornado. 6/4/05 Hiawatha storm ... also, I do not recall serious CG activity until after the tornado was completely finished (power flashes were prominent, on the other hand). There are many other examples I can think of as well, but these just stand out. However, when a rotating, tornadic supercell approached St. Joseph later that evening, there was lightning on a scale that I cannot describe (in fact I just happened to write about it in my blog early this morning). There was absolutely incredible bright blue and green strokes striking the ground literally all around the meso, reaching high into the updraft. I described as it having an 'other-worldly' feel to it, because it was almost beyond reason, and beyond anything I've personally experienced (and I also think a bolt managed to sideswipe my car, as my electrical system is now fried). This was true strobe, CG lightning - and it was amazing.

My own theory goes something like this ... it takes a while for a 'new' storm to reach a level of maturity that allows for high electrical output. I do not believe we see the truly noteworthy electical activity until the storm has managed to stabilize a seriously strong updraft, pushing an anvil further away from the tower and covering more and more area as the updraft forces material higher and higher. A tornado is still possible relatively early in this process, before the electricity starts trying to discharge, because all of the ingredients necessary for a tornado are there, but not necessarily for high electrical activity yet. The Hallam storm did begin to produce a lot of CG strokes, etc., but it was only after the storm reached an appropriate age for it ... the same was true of Hiawatha and others (Falls City/Mound City supercells in July, '02 are another good example). By the time the St. Joseph storm approached the Missouri River, it had already been around on radar for some time, as I had been watching it. It took longer for that storm to begin indicating serious ground level rotation, and by the time it did the ingredients for both a tornado and high electrical discharge were both there at the same time, thus allowing for the activity in and around the mesocyclone (and I'm assuming all the way up the updraft tower into the anvil at that point as well).

In general, I feel like high electrical activity in tornadic supercells is largely the result of other mechanisms at work, but as chasers we notice it only once the storm has reached a very mature stage - I really think that chasers who arrive late after the rotation has already been around for a while are the ones to continually notice this, though, and I think that had they been with the storm from the point of initiation, they would also notice that it takes a while for the electricity to get in gear. I've seen threads by those who understand the processes of lightning production a lot better than I do of course, but I would like to take some time at some point to explore this conversation more (again), simply because I cannot for the life of me justify a true connection between high CG/CC activity and tornadogenesis. I suspect that has more to do with ice particulate separation that Jeff S. mentions, and the ability of a mature supercell to successfully accomplish this. Hope this makes sense -
Jeff, I do have one notable exception to your 2003 no-lighting supercells. On May 9th, 2003, I intercepted a supercell near Union City that was to produce tornadic activity in the OKC and Stroud areas. I approached it from the east as darkness was falling, and I remember this cell as having some of the most continous and intense lightning I have seen to date.

I don't remember there being a lot of CGs, but the lightning within the storm was strobelike in its intensity. I don't believe it had dropped a tornado before that time, but it did go on to become a long-lived multiple tornado producer.
Jeff, I do have one notable exception to your 2003 no-lighting supercells. On May 9th, 2003, I intercepted a supercell near Union City that was to produce tornadic activity in the OKC and Stroud areas. I approached it from the east as darkness was falling, and I remember this cell as having some of the most continous and intense lightning I have seen to date.

I don't remember there being a lot of CGs, but the lightning within the storm was strobelike in its intensity. I don't believe it had dropped a tornado before that time, but it did go on to become a long-lived multiple tornado producer.

Here is a video of this night as it was moving over the Bristow area. A few cg's but mostly behind the wall cloud but more in the way of flashes during it's passing
Wow. The lightning really died down in between where I first intercepted and Bristow. I didn't notice much lightning after the storm became tornadic, but that might be because I was in the rain core. :)
My most vivid experience with anvil zits and the soft continuous thunder they produced was during the 10/9/2001 outbreak - I can recall hearing soft continuous thunder throughout much of the life cycle of the Foss Lake tornado. Although the supercells were no longer tornadic once they moved into the OKC area after dark, we did have a vivid display of anvil zits here in Norman that evening as one of the supercells moved just NW and N of here.
Since my ongoing Master's research involves heavy analysis of lightning activity in convective storm environments (including a supercell case 25 April 2005 Dallas-Ft. Worth), I am excited to see this topic brought up by Dan.

The one consistency between supercells is that their lightning activity is as unique as the weather they produce. Unfortunately, there are no lightning behavior signals that can be related across the entire domain of supercells to predict tornadic activity. In a supercell, the favorable location for total lightning (the in-cloud breakdown of both IC and CG flashes) is in the downshear reflectivity gradient extending toward the FFD. The vigorous updrafts of a supercell and ambient favorable shear environment are the ideal set-up for charge separation and large quantities of IC flashes. Most of the IC occurs in the mixed phase region between 6-9 km so it may not be audible or even visible to the eye thanks to intense hail/rain precipitation cores in this vicinity. The "anvil zits" that Dan has noted are likely interaction between the elevated negative charge region (~ -15 C) and the upper-positive charge region (~ -35 C) due to a vigorous updraft that has separated them sufficiently to initiate nearly vertical lightning in the immediate updraft region.

A total lightning signature known as a "lightning hole" has been observed by researchers in conjunction with many tornadoes. Basically, the updraft of the supercell becomes so strong that ice crystals and graupel are evacuated from this region leaving a "hole" in the total lightning imagery, since electrical breakdown can not occur without the presence of these charge bearing hydropmeteors. This "lightning hole" has corresponded well with the location of reported tornadoes in these instances. Unfortunately, this signature is not there in all tornadic cases (e.g. the 25 April 2005 Dallas event, although it was a marginal tornado case). The link below is an article regarding the 2000 STEPS project where these "lightning holes" were first observed.


One last thing of note, CG activity tends to lag the total lightning activity by 10 to 20 minutes as the IC flashes favor an elevated charge regime (resultant from strong updrafts) and CG activity favors lowered charge regimes (downdraft becomes dominant), so a cyclical behavior of the IC and CG flashes is noted often times with supercells as they evolve over time.
Supercell Lightning

When I lived in the Ohio Valley, I noticed a different yet equally strange lightning occurrance with tornadic and pre-tornadic supercells.
There would be a single discharge or pulse of colored lightning that would last as much as two seconds and change colors from luminescent blue-green, to green, to red, to green and back to blue-green. It was as if someone simply turned on a light that changed colors then swiftly faded out.
During the rare occurances of this type of lightning, intense storms were in progress and included brief, minor tornado damage reports.
Lightning frequency in tornadic cells

Dan, I'm so glad you brought this topic up! :D

I was talking about this very thing with a friend of mine not long ago.

Over the years I've seen a frequency to the lightning/flashes within a storm when rotation is associated with it. It's good to see that others have noticed this as well. I've tried to explain it to a few fellow storm chasers and it was hard to explain.

One example:
Last summer I was chasing a storm and noticed that frequency of lightning pulses in this one area of sky. It was like the action I'd seen with storms with rotation in the past. I was out of range for my cell phone coverage and hung back a bit from the storm(it was at night). When I finally connected to Joe Lawton, I found out that I was right the cell had a tornado warning with it.

I've seen this lightning activity with all rotation I've seen/been near to. Once again good to see others noticing this.