Tornadogenesis and the role of the RFD

Joined
Oct 15, 2015
Messages
50
Location
Calgary, Canada
Hey folks,

First of all, I'm new to this forum and am excited to be a part of it. To my knowledge, we don't have anything like this in Canada (and am not sure how many Canadians use it), so I'm stoked to be able to discuss severe weather here and hopefully learn some things from you folks, especially from experienced chasers and meteorologists.

*Be advised that this post is LONG, so please, read and respond at your discretion :)

I understand that the topic of how tornadoes form is still in many ways poorly understood, however my attempt here is to solidify the current knowledge we do possess on the subject today. Some questions remain outstanding for me (and maybe they do for everyone). Four sources of info that I have been referencing lately represent both a meteorological perspective and a storm chaser/spotter's perspective, and include:

How to Make a Tornado, by Markowski and Richardson
- http://www.weatherwise.org/Archives/Back Issues/2013/July-August 2013/torando_full.html

What we know and don't know about tornado formation, by Markowski and Richardson
- http://scitation.aip.org/content/aip/magazine/physicstoday/article/67/9/10.1063/PT.3.2514

Skip Talbot's youtube video "Storm Spotting Secrets"
-

Advanced Storm Spotter Training Webinar, from NWS Norman
-


1) With reference to the scholarly articles, the gist of their message is that environments conducive to strong tornadoes tend to have two things:
- Enhanced low level shear, which increases the strength of the mid level meso, which thereby increases low level "suction," that of course is crucial for low level vorticity stretching
- Increased low level moisture, which leads to less evaporative cooling from downdrafts, which leads to relatively warmer downdrafts (though still possessing baroclinicity), that aren't so negatively buoyant that low level vorticity cannot be ingested into the updraft

However, no specific mention is made of an "RFD" as opposed to an "FFD." In fact, from the images that are used, it almost seems to be implied that the low level vorticity that is generated along the interface of rain-cooled downdraft air and ambient (inflow) air originating from the FFD is responsible for tornadogenesis. This is evidenced by the fact that it shows air descending down from right to left from the main precipitation cascade to beneath the wall cloud. If this is the case, what, if any, role does the RFD play in tornadogenesis?
In other conceptual models I have seen for supercell wind fields, it looks like tornadoes form at the occlusion point between the RFD (wrapping into the hook) and the inflow air.

2) Skip's excellent resource for storm chasers provides a way to "landmark" some of these processes in being able to more precisely locate where a tornado may form in a supercell thunderstorm. He does make mention of the RFD as being a process that "feeds and drives a tornado," and shows that the RFD punches into the updraft base, creating a "horseshoe" shape, characterized by a clear slot that often forms from RFD air. (See 5:50-10:50).

On a reflectivity scan of a supercell, the "hook echo" is the visual manifestation of the RFD, is it not? How then can it be representative of drier, subsiding air? I know that there can be dry and wet RFD's (the latter of which often being associated with HPs), and that supercell mode can change within the same storm, however is there always precipitation reaching the ground beneath the hook echo? One would think that precip falling into drier air would then have a greater evaporative cooling potential, thereby increasing the negative buoyancy of that air, potentially causing tornadogenesis failure. But this clearly doesn't follow in many cases.

3) In the NWS resource, from 1:02:30 on, the origin of the RFD and mechanisms for tornadogenesis in general seem vague (perhaps because an unpacking of this concept isn't the intent of this resource), but some claims are made that seem to be contrary to the above. For example, at 1:03:54, the speaker says "you really need a warm, unstable RFD to stretch that vorticity in the low levels," which is clearly at odds with the idea that the upward directed PGF from the strength of the mid level meso is what is responsible for vorticity stretching. Moreover, in their hypothesis on the origin of the RFD, they claim that there is relative high pressure in the mid levels and strong low pressure in the low levels, that is responsible for a sort of "downward directed" pressure gradient force, that is in turn responsible for subsidence of RFD air. Of course, this too seems to be at odds with the above theory. (I have heard alternate explanations that RFD air is FFD air that has been wrapped around the mesocyclone and modified by mid and upper level air that has been forced to descend after colliding with the back of the storm).
The resource also claims that RFD air causes the clear slot, and that it is "hot and unstable" (though presumably not so "hot" as to not contribute to baroclinicity).
Is there any generally agreed upon hypothesis as to the origin of the air the RFD? And if it's properties are different from those of FFD air, are there in fact 3 different air masses within the storm scale vicinity of a tornadic supercell (ambient/inflow, FFD, and RFD)?

Congrats if you've made it through this and have seen my 3 general questions through all the chatter. Again, I realize that there may be no agreed upon explanations of the questions I've presented - though I do seek clarification for apparent inconsistencies in current scientific knowledge. The questions were, again, in sum:

1) To what extent does the RFD (as opposed to FFD) play a role in tornadogenesis?
2) Is there always precip reaching the ground beneath a hook echo, and can an RFD associated with an echo be simultaneously associated with a clear slot?
3) What is the origin of RFD air, and are there three distinct air masses in the vicinity of most tornadic supercells?
 
This matter is extremely complicated and I'm not even going to attempt to address/answer the 3 questions. However, you may like this ongoing research utilizing a high-resolution simulation of a tornado by Leigh Orf:
 
I did NOT read your whole post, Kyle, but here is a brief answer to your questions: You don't get a tornado unless the storm-scale rotation gets transferred right down to the ground (which is one reason low LCL's are "helpful"). Thus, the RFD can be the last straw that is needed to get that rotation to the ground.

Hook echoes occur in the same place as the RFB, ergo there is not always (substantial) precip there. I wouldn't say an "RFD (is) associated with an echo". That's just sort of an apples/oranges thing, but most of the time, if there's a tornado there is also an RFD and a clear slot. It's all part of the package, along with large hail, damaging winds, etc. Remember, though, no two storms are the same. I'm only talking about supercells here.

I'm no expert on RFD's, but they sure do seem to originate in the upper levels of a storm, perhaps driven downward from being trapped between the upper jet and the back-shearing anvil in many cases. Three distinct air masses? Again, every storm is unique, but yeah, pretty much.
 
1) To what extent does the RFD (as opposed to FFD) play a role in tornadogenesis?
2) Is there always precip reaching the ground beneath a hook echo, and can an RFD associated with an echo be simultaneously associated with a clear slot?
3) What is the origin of RFD air, and are there three distinct air masses in the vicinity of most tornadic supercells?

I'm not going to take the time to provide highly detailed/discussed answers to these questions because I don't know the answer for sure and I don't have the time. With that said...

1) The RFD very likely plays a crucial role in tornadogenesis because there are few, if any, cases where a tornado formed from a mesocyclone in a supercell without an RFD present.

You should also consider the "vortex arches/rings" theory by Paul Markowski. The role of the RFD in that theory is to shove the vortex rings down towards the surface and help erect it into the vertical. AFAIK, it is generally accepted that the rolling air tubes that produce both the mid-level and low-level mesocyclones do not get tilted enough right at the surface to produce the tornado scale rotation. Something else has to come along and shove the arcing vortex tube all the way to the ground. The theory states that the RFD is responsible for this. See this short paper, specifically, Figure 2, for an illustrated version: http://www.meteo.psu.edu/~pmm116/pubs/2009/MR09ATMOSRES.pdf

2) This question is somewhat unclear, but here's my response to what I think you're asking: A hook echo doesn't necessarily mean there is an RFD present. I believe it has been shown that precip can just fall out along the flanks of the updraft, and some of it may rotate around the updraft and reach the ground on the backside of the storm. You can call that a RFD. However, descending air warms. A developing downdraft does require falling precipitation (yes, you can get weak descending motion from synoptic scale processes, but that's clearly not relevant in a supercell). Occlusion downdrafts also occur, but those are smaller in scale and are associated with a vertical perturbation pressure gradient force. I'm not sure how relevant occlusion downdrafts are to tornadogenesis. Anyway, whether or not precip is falling within a downdraft depends on how much moisture is available in the downdraft. If available ice and water are completely scavenged, the downdraft can reach the ground with no falling precipitation.

A "clear slot" is a human observation in that it's a perception to your eye. Just because there is a "clear slot" doesn't mean there is no falling precipitation in it. You can't see individual precipitation particles from hundreds of yards or even kilometers away. There can still be enough particles in there to result in some reflectivity, however. So, yes, an RFD can show up on radar with reflectivity while also appearing to the human eye from a distance as a clear slot.

3) No one knows for sure exactly how the RFD forms. Observational studies don't have an easy way to measure anything up there. There are some published articles in the literature on this topic. I suggest
http://journals.ametsoc.org/doi/abs/10.1175/1520-0493(2002)130<1692:DSTOWT>2.0.CO;2 and
http://journals.ametsoc.org/doi/abs/10.1175/1520-0493(2002)130<0852:HEARFD>2.0.CO;2

and perhaps check later papers that cited these papers.

It stands to reason that there may be three separate air masses in supercells. However, thermodynamically speaking, the FFD and RFD can be quite similar to each other. Typically there's more of a kinematic difference (in wind speed and direction).
 
Thanks for the input, gents. Caleb, that simulation was very helpful for visualization of the processes related to this discussion.

Jeff, those are some great resources and I'll add them to my reading list. Also beginning to work through Markowski's Mesoscale Met publication, but I figured this discussion would be a good primer. Your answers also addressed my questions nicely.
 
A couple things:

You don't get a tornado unless the storm-scale rotation gets transferred right down to the ground (which is one reason low LCL's are "helpful").
I think it's generally thought that low LCL's are helpful because low LCL's imply high boundary layer relative humidity, which reduces the propensity for the RFD cold pools to be too cold. Cold air is hard to lift (you know that already, I'm sure ;) ). There's some evidence that at least some of the air that tornadoes ingest is RFD air, and RFD parcels that are too negatively buoyant would disrupt the circulation of any developing tornado.

With that said

The clear slot is the visual manifestation of the RFD (Lemon and Doswell 1979 supercell paper: http://journals.ametsoc.org/doi/pdf/10.1175/1520-0493(1979)107<1184:STEAMS>2.0.CO;2). It's "clear" because the descending (drying) air has evaporated all the cloud droplets. The smaller a drop is, the higher its surface-area-to-volume ratio, which means it's easier to evaporate. So the cloud droplets go first, then the small raindrops, then the large rain drops evaporate last. But out of those three, the radar is most sensitive to large rain drops, which you can't see with your eyes from a distance. It's even more sensitive to hail, which is also hard to see with your eyes and is hard to evaporate. So that's how you can get high reflectivities in a "clear" slot.

It's also worth noting that parcel theory doesn't seem to apply well in the RFD. The downdraft gets diluted too much, and the dilution increases when shear increases. This violates a major assumption in parcel theory. These were some of the main findings of a cool Gilmore and Wicker paper from 1998: http://journals.ametsoc.org/doi/pdf/10.1175/1520-0493(1998)126<0943:TIOMDO>2.0.CO;2

Matt Parker (of NC State) gave a presentation at OU a couple months ago, and part of that was research published earlier this year with Johannes Dahl (here: http://journals.ametsoc.org/doi/pdf/10.1175/MWR-D-14-00310.1). It essentially demonstrated that the presence of a downdraft and vertical shear were sufficient conditions for developing *some* vertical vorticity near the surface. This vorticity is not necessarily of tornadic strength, but it is in a location where it could easily be stretched to tornadic strength by an updraft, should there be one (there wasn't in their experiment). That stuck with me because those conditions are met almost by definition in every supercell. Which means we might be asking the wrong question. Instead of asking "what do the storms that produce tornadoes get right?" the question might be "what do the storms that don't produce tornadoes get wrong?"

I'm currently working on a literature review of some of this very material for my general exam. I might unearth some more things that are of relevance, and if I do, I'll post them here.
 
Last edited:
Very good discussion!

This video from Dr. Markowski to the NWS sums up the great, well-documented 2009 case study. It goes in depth on tornadogenesis a little beyond the aforementioned articles (especially visuals).

The Leigh Orf video Caleb mentioned is another great visual of tornadogenesis and can help visualize origins of RFD air.
 
This is a great discussion topic, and the video link above is fascinating - everyone should watch!

My own personal observations from watching developing tornadoes is that there is often a surge of low-level air from within the forward-flank core just prior to development. At times, visually, it's hard to distinguish this from the RFD - but in the case of the El Reno tornado, I noted extreme motion from 'out of' the main core of precip in the few minutes before the tornado developed. In these few minutes the wall cloud, initially ragged, quickly became well-formed, and started to exhibit rapid cyclonic motion. Perhaps this surge is driven by a sudden downburst/microburst close to the low-level mesocyclone. These 'descending rain cores' have previously been areas of interest for study.
What is also interesting is that the current thoughts suggest most if not all of the low-level inflow into the base of the tornado is air which has already been 'processed' by the storm. 'Warm' inflow into the tornado seems to occur somewhat aloft, after the air has been lifted over the forward-flank boundary.
 
There are others who know a lot more about this than I do, but to elaborate a bit on what Tim said above, one important point to keep in mind is that the temperature in the RFD can vary quite a bit under different conditions, and that plays a big role in the effect of the RFD on tornadogenesis. If the RFD is cold, it can indeed undercut the circulation and make tornadogenesis unlikely. OTOH if it is warmer and therefore more buoyant, it can help bring the circulation to the ground without undercutting the strength of the rotating updraft. In this case it can and often does play an important role in tornadogenesis.
 
The video @Taylor Stone posted was excellent.

The one thing I am a little unclear of from that presentation is the mechanism that causes the vorticity vector to lift off of the trajectory of a parcel descending to the ground (analogy of a plane landing with nose up attitude as it nears ground). We know that in order to get a tornado to form that we need near-ground, vertically oriented vorticity, that is then stretched and intensified due to dynamic "suction" from the mesocyclone. This vorticity is believed to be generated by horizontal vorticity that forms as a result of baroclinity along the interface of ambient/downdraft air, that gradually tilts into the vertical (along its trajectory) near the ground. Is there an explanation as to how this happens? Put another way, why wouldn't the vorticity vector (originally oriented along the path of the trajectory) remain in sync with the trajectory all the way to the ground?

On a side note, I like that he used the term "baroclinity" in the presentation. I've only ever heard/seen the mouthful word that is "baroclinicity."
 
Also, since the mechanisms that cause the low and mid level mesos to form are different, apparently they are not always "coupled" to each other. One would think the low level meso would naturally merge upward with the mid level meso, especially due to the strong VPPGF that is induced by the latter, and would presumably cause subsequent intensifying of the former. Has anyone ever observed/heard of a tornadic low level meso that was decoupled, or independent of the mid level meso?
 
The RFD is simply a vertical standing wave of entrainment. To me, a non-scientist, the RFD is half the yin and yang. When seeing an overall meso, as seen on radial vels as apposed to in the clouds, the RFD is half the circulation. The tornado just in the center, often on the head of the comma created by this entrainment of the RFD. Now if the RFD is cold or dry, it has a bad theta-e and chokes off the tornado or pre-tornado circulation. Although if the RFD comes out with a good, high, theta-e, then look out. This RFD air can be as buoyant or moreso than the warm sector air east of it. Then tornadoes can occur, or get very very big.

We are much like surfers. We paddle out to where the waves will break. Bragging about life and laughing at the possibility of rocks or sharks. The the wave comes and we paddle like hell to catch it. Ours is just vertical, and 'the tube' has a much more intense meaning.

I'll let the more science minded above lead you to where the RFD comes from. But I still like the idea that it is just entrained air, then modified by passing precip through it. High SRH leads to helically spiraling updrafts. That turret look. This is the storm entrainning air which does not just 'mix in' to the updraft. Instead it is hated and shunned by the updraft. It's pulled in to the updraft and since heavier, sinks. On the way down, you can pass precip, hail and whatnot, through this entrained air, raising its theta-e. If you just put precip IN to this air, then cooling will offset any evaporated H2O. But you can cheat and pass the precip through it, allowing added H2O without 'all' the associated cooling.

We need all chasers to have 'accurate' theta-e measurements on their mesonets sent to NWS. :)
 

Here is another presentation by Leigh Orf which is a bit more detailed than his first presentation. Some of his findings suggest much of the air being ingested by the tornado comes from the RFD. Another feature he discusses is the SVC (streamwise vorticity current) which is often seen as a lowering usually immediately north or northeast of the tornado. The 3D animation gives one of the best visualizations yet on how the air flows throughout a tornadic supercell.
 
Back
Top