A Case Study in Weather Forecasting:

Supercells



This case is told by Mr. David Etheridge, a senior civilian forecaster at the NASP METOC and a true "weather lover." He has been forecasting since 1966, including a tour in Viet Nam and service for the Strategic Air Command. He also served for a time as Instructor at the USN Aerography School. He has logged over 54,000 hours at forecasting or forecasting-related tasks.



Severe weather events such as "Supercells" can develop, mature, and dissipate in a classic pattern illustrating a standard scenario, but can depart from the classi pattern in other respects such as the time of day or time of year.  In such cases, the Forecaster needs to notice salient features, sometimes under conditions of limited data, and reason hypothetically.  Features that are notifed and are suggestive of developing severe weather need to be monitored, but especially they need to be explained, and never  igonred.  This case study also illustrates what it means for the Forecaster to "beh his brother's keeper."


 

 

Event  and Comments

EVENT TYPE

 

Time

Early 1998

February or March

 

 

 

Midwatch

Nothing was said at the watch change--no guidance to look for storms.

I had continuity (on-going understanding of the weather over previous days). 

The immediate goal was to TAFs and other routine stuff.

 

Action

5:00 PM

When I started the watch at 5:00 PM the system was not appearing on the radar.

At the start of the watch there was no weather going on at all.

 

Observation, Situation Assessment

 

Before the Mobile radar painted it, we knew only that there was a SW flow and clouds down over the Gulf south of New Orleans.

It was off the scopes. But on night shift you have the slack. 

You get few phone calls, etc. and you can take the time to look around.

I noticed this when it was way way out.

This fit a standard scenario--South-west flow over a water source. 

You will not take your eyes off it. Since it was far away you could look upstream.

 

Observation, Situation Assessment

 

I monitored the radar

Supercells can last a long time. 

One I followed went through six states from Texas to out over the Atlantic. 

It spawned a tornado and did lots of damage.

When it was way down a novice might have seen it, but might have said it would fall apart, or might not have paid any attention to it. 

It was a small speck way down. 

A novice might have not even looked. 

But if it maintained for an hour or so and got within 50 miles out he'd have started studying it.

 

Action, Observation, Situation Assessment

6:00 PM on


Mobile radar started to paint it.

It was about 120 miles out, at the extreme range but you could see the tops--upwards of 40,000 feet.

At this point, to a relief officer I'd have to say don't take your eyes off of it. 

You may have to issue a warning as there may be waterspouts, tornadoes, severe weather.

The alternatives were limited in terms of what I could do.

You could look at bouys and GOES. 

I did look at those. 

Bouys were not useful as this was a solitary cell. 

It would have to run right over a bouy and that would be unlikely in the Gulf--there aren't many bouys. 

The best data source would be direct observations--people out on the oil rigs. 

But this was at night so all they could see would be lightning or back-lighted lightning.  

On GOES you would see the upper-level cirrus difluence.

 

Observation, Situation Assessment

9:00 PM

 





After one hour I knew it would qualify as a supercell.

Once it was at 120 miles out there was no doubt. 

The decision was easy.  

Its structure was a "solid state," meaning it maintained itself.

They usually dissipate but if they maintain that shows it has the dynamics of supercells.

A novice might have missed it as a supercell for an hour or so but after two hours he would have found it. 

Besides, the Weather Channel would have noticed it (see 11:00 PM, below).

 

Decision

10:00 PM

NEXRAD showed a hard core in the right rear quadrant.

It had good movement.

The heaviest echoes were in the right rear quadrant. 

It looks like a turkey walking away from you (see the above diagram) and it seems to "waddle" in the time lapse at its back side where the hard core is.

At long range you have time to notice things like this. 

It is a luxury. 

When systems are close-up you are busy doing your warnings.

 

Observation, Situation Assessment

 

You know this from experience.

I've watched these for 25-30 years, using about 6 different radars (USAF and USN) over the years.

NEXRAD does storm tracks every 15 minutes. 

You can also do it yourself by laying down marks.

You can put marks on the radar screen. 

It is a good aid. 

You can put your own dots down with time marks. 

We do this for hurricane eye walls. 

I did it now and then to see which way the cell was going and see where it would go.

The old radars gave just color shades. 

You could get an idea of the vertical configuration but they were not nearly as good as NEXRAD. 

NEXRAD is a wonder of technology.

 

Decision

 


 

It was 100 miles SW of Pensacola, over the mouth of the Mississippi River.

It was a big cell.

It was moving NNE.

It was moving at about 20 miles per hour.

To be a supercell it has to be 10 or more miles in areal extent.

This one was 30 - 35 miles across.

 

Observation, Situation Assessment

 

If anything lasts more than one hour, any storm, and has the dynamics to stay together (indraft at the bottom and outdraft at the top) then it can develop into a supercell.

Most storms  last for an hour or less.

If they have a tilt and high upper-level winds and they keep moving then they can keep building.

But if they are slow they just choke off.

But any time you have SW flow, or if you have vorticity,  something can gin up.

C-school covers this. 

They do severe weather scenarios with surface and upper-air analyses. 

You should not be blind-sighted.

 

Decision

 

I kept monitoring the radar.

If there had been a change in size--shrinking to 20 miles in areal extent or the tops getting less tall or reduced dBZ from 65 down to, say, 55--all this would imply that it was dissipating, that it had dropped all its moisture. 

But you'd still watch it.

It might not spawn tornadoes, but still could involve high winds and hail. 

Even a novice would keep watching it. 

It could still have destructive potential since it was still upstream at this point.  Once it has your attention you bird-dog it until it goes away.

 

Action, Observation, Situation Assessment

10:00 - 11:00 PM

I knew it would hit close by and would effect our Areas of Responsibility.

I extrapolated its track, but NEXRAD did that too.

NASP would not be in danger but I started to worry about Eglin and Hurlbert Fields.

 

Decision

11:00 PM

When it crossed the mouth of the Mississippi, the Weather Channel said "Look at this supercell!"

They showed NEXRAD and said nothing about where it was going. 

It was way out over water and hours away from affecting anybody. 

The NWS hadn't issued anything but the Weather Channel noted it just as an interesting feature. 

There was no other weather anywhere for them to talk about.

 

Observation, Situation Assessment

 

We'd been looking at for over an hour and a half.

I had no alternative courses of action.  I'd have not called Tallahassee.  

And Norfolk would have been no help either. 

NWS severe weather work is done at Kansas City. 

We don't call them. 

It would be good to be able to discuss sever events with them but they would never go for it.

 If every Tom, Dick, and Harry called them it would pull them off their job.

 

Action

 

It was a bad system. 

Bad systems like this one are usually in February or March.

It fit the standard scenario.

"Bad" means severe--winds, hail, possible tornadoes.  Bad is when all the dynamics line up.

February and March makes for bad Especially when it comes out of the South-west.  It has warm air and then hits a moisture source.

 

Decision

 

Slidell and Mobile radars were getting good reads on it.

The key features were the hard core in the lower quadrant and high tops. 

The storm was stable and was moving consistently.

 

Observation, Situation Assessment

 

It kept coming NNE.

 

Observation, Situation Assessment

 

I kept extrapolating the track via NEXRAD.

I watched it sweep by sweep.

The tilt shows that the indrafts do not converge on the downdrafts.  If they hit, they cancel. 

The vertical NEXRAD profile shows the tilt. 

Tilt implies that it is a really bad storm.

(E did a diagram).

I stood up and did a 360 looking at the Wall of Thunder and the NEXRAD.

 

Action,

Observation, Situation Assessment

From 6:00 PM  on





Bad cells tend to turn to the right but they can sometimes turn to the left.

They tend to turn to the right since the indrafts come in from the lower right flank causing new cells to form in the lower left flank as the older cells dissipate. 

The turning is virtual. 

It is not that the system turns, it is a matter of where cells dissipate and new ones form.

 Systems will turn toward the left only if there are upper-level winds driving in, winds up between 500 and 700 millibars do the steering.  The last case I saw of one turning left was a storm moving east to west across the panhandle. 

It turned left and moved over Pascagoula. 

I never had time to do much analysis on that one.

The other dynamic is in older storms. 

They divert the upper air and can push the newer storms. 

One of them can hook around the other. 

Sometimes they get caught up in each other's flow and merge and turn into really big storms. 

The movement can be "apparent," though. 

It is never easy. 

You never have all the data.

You have to make assumptions.

That's where experience comes in.

You know to look for these sorts of things.

 

Decision

 


 

If it is upstream of you, you are not going to take your eyes off of the radar.

At about 11:00 PM I could tell from the storm track that I should be able to see lightning. 

The system was about 120 miles out. 

I decided to go out and look with the Observer. 

I often go out with them to verify their forms and codes (cloud types, lightning types, etc.).

The lightning was almost continuous. 

I could see it behind the tree line outside to the south.

You can tell by the lightning how bad it is. 

If the lightning is continuous that implies it is a supercell. 

Also, you can hear it once they come up to about 30 miles. 

You can get inter-cloud lightning overhead. 

It is about five miles up in the horizontal part of the storm.

 

Decision,

Action

11:00

Once it tracked to the 100-mile mark I thought it would hit east of NASP, over on Pensacola Beach to Navarre.

I knew it would hit at about 3:00 AM.

 

Decision

 

I knew I'd not have to put out a warning for NASP

Whiting Field was on the same bearing as NASP given the cell's track. 

If it would miss us, it would miss them. 

Otherwise I would have had to be careful. 

No errors were possible at this point.

If it were a flying day I'd have had to put out a warning--there could be hail, waterspouts, etc. 

I'd have closed off some of the flying routes. 

Search & Rescue would talk to us before they flew if they had to fly.

 

Decision

 

I kept monitoring it.

 

Action

 

When it got 40 - 50 miles SSW of NASP I realized it would track 20 miles E of NASP, right about at Hurlbert Field.

 

Decision

1:00 AM

I called Hurlbert Field.

I did so out of professional curiosity. 

Late at night you aren't busy so this was a courtesy call to discuss the system--How bad Hurlbert thought it might be.

 

Action

1:00 or 2:00 AM

We'd been watching it for 6 hours by then.

 

Action,

Observation, Situation Assessment

 

They had a young forecaster there, just out of Italy, new to the area and hadn't worked any severe weather.

I turned him into a bird dog at just the right moment.

I assumed he'd contact Eglin too since both are USAF and watch each other's TAFs. 

They might even have a hotline.

 

Observation, Situation Assessment

 

I asked if he was aware of the supercell heading toward him?

He said, "Huh?  What?"

Others might not have called Hurlbert. 

We assume that others are doing their jobs. 

If I hadn't called, he'd have been preparing to do TAFs and Monday through Friday overnight stuff to midshift--the next day's weather for briefing their staff. 

He was within 2 - 3 hours of getting hit and he hadn't even been watching it. 

It would be nice for there to be a checklist for NASP, Eglin and Hurlbert to do more information sharing and weather discussion, especially on the night shift (on day shift there's lots of data floating around). 

We can call Mobile but we don not do any local discussions. 

We only talk to Tyndall, Eglin or Hurlbert if they call us for winds, waves, etc.  For stuff over water they USAF bases tend to have a blind eye.

 

Action

 

I said, "Well, you got a supercell coming up on you. In about 90 minutes you'll get some business.  You should think about putting out severe weather warnings for your Base and claimancy."

 

Action

 

I assumed he got busy--doing his checklist--even before USAF Weather Center would have put out a Point warning, which would trigger his checklist.

So I did not bother him after that.

I never went back to see what he did.

 

Decision

 

He probably made some phone calls.

They would get busy.

They may have issued a Severe warning or even a tornado warning.

 

Decision

 

It was east of NASP making landfall.

I was able to watch it through the front windows. 

There was a continuous light show behind the trees as it went south-east to east of us. 

There was audible thunder.

On NEXRAD I saw some fast winds and a heavy gradient.

 

Observation, Situation Assessment

3:00 AM

I watched the supercell go right over Hurlbert Field on the radar

It was a heavy cell, but too far from the Mobile radar to be able to see any tornado signature (for that you ideally need to be within 60 miles of the radar).

I was concerned that it might spawn tornadoes. 

You always forecast for the worst possible situation. 

I'd have issued a Severe-1 if it had been closer to NASP.

We'd have seen a signature. 

Eglin has their own radar and Hurlbert probably piggy-backs on it.

 

Observation, Situation Assessment

 

It was 25 - 30 miles across.

 

Observation, Situation Assessment

 

We got  lots of lightning here at NASP and we were about 10 miles off the NW edge of the supercell.

 

Observation, Situation Assessment

 

As I learned later on the radio (about 6:00 AM),  a tornado was reported by the Highway Patrol in Crestview, near I-10 at about 4:30 - 5:00

 

Observation, Situation Assessment

4:30 - 5:00 AM

At daylight the supercell was NE of us, tracking up toward Crestview.

 

Observation, Situation Assessment

 

At the watch change,  I just said that a cell had gone through and that no warnings were in effect here at NASP.

It was a migratory system and could not have backed up. 

The surface winds in winter move and lock you into the flow. 

Nothing could have moved back up toward NASP.

Upwind to the south-west of NASP nothing else was appearing.

 

Action

5:00 AM

We ended up watching it a total of about 9 hours.

There was no stress; it is just doing your job. 

It was not a routine situation since it is not usual to get a heavyweight supercell at midshift. 

Ninety percent of the time severe weather is rare for that season. 

But this was a standard scenario in terms of the storm development and dynamics, but not standard in terms of the time of year and time of day of the storm. 

Fast moving cold fronts coming from the west usually determine our winter weather--storms and small lines of storms. 

A big cell developing in a south-west flow is rare for winter in the Gulf region.  In terms of the time of day, in the Midwest peak heating generates severe weather in the time frame of 5:00 PM to midnight.

 

Action, Observation, Situation Assessment

 

There was no time pressure--I saw the system whenit was well upstream.   It took little time for me to make the decisions I made.  There was no stress; it is just doing your job.    The decisions I made were the best.  Experience was the key factor.  The only mistake that I could have made was if I had stopped monitoring the system.

 

A novice would have done pretty much the same, but may have caught it later than I did.

 

The only key feature of the system that would have made a big difference was the characteristics of a supercell.  Any change would imply it was disipating.  On GOES you would look for a change in the "V" pattern to the difluence aloft.  If it were dissipating the difluence aloft would have thinned out, implying that the cell was weakening.  But in this case the outflow aloft stayed heavy.

 

It was not a routine situation since it is not usual to get a heavyweight supercell at midshift.  90 percent of the time severe weather is rare for that season.  But this was a standard scenario in terms of the storm development and dynamics, but not standard in terms of the time of year and time of day of the storm.  Fast moving cold fronts coming from the west usually determine our winter weather--storms and small lines of storms.  A big cell developing in a south-west flow is rare for winter in the Gulf region.  In terms of the time of day, in the Midwest peak heating generates severe weather in the time frame of 5:00 PM to midnight.

 

 

 

 


Decision Requirements

 

Cues and

Variables

 

 

 

 

·        Skywatching of lightning and thunder can provide clues as to whether a storm is a supercell.  Lightning can be observed even from great distances.

·         

"I could tell from the storm track that I should be able to see lightning.   The system was about 120 miles out.  I decided to go out and look with the Observer.  I often go out with them to verify their forms and codes (cloud types, lightning types, etc.). The lightning was almost continuous.  I could see it behind the tree line outside to the south. You can tell by the lightning how bad it is.  If the lightning is continuous that implies it is a supercell.   Also, you can hear it once they come up to about 30 miles.  You can get inter-cloud lightning overhead. It is about five miles up in the horizontal part of the storm."

 

·        Supercells can last for a long time, even days.

 

"If anything lasts more than one hour, any storm, and has the dynamics to stay together (indraft at the bottom and outdraft at the top) then it can develop into a supercell.  Most storms  last for an hour or less. If they have a tilt and high upper-level winds and they keep moving then they can keep building. But if they are slow they just choke off. But any time you have SW flow, or if you have vorticity,  something can gin up."

 

·        Upper-level difluence seen in GOES is an important indicator of supercell development.

·        Supercells are indicated by a "hard core" in the lower quadrant as seen on NEXRAD.

·         

"NEXRAD showed a hard core in the right rear quadrant. It had good movement. The heaviest echoes were in the right rear quadrant.  It looks like a turkey walking away from you (see the diagram) and it seems to "waddle" in the time lapse at its back side where the hard core is."

 

 

 

 

 

 

 

 

·        A change in size of a supercell is a key feature in terms of developmental phases.

"Shrinking to 20 miles in areal extent or the tops getting less tall or reduced dBZ from 65 down to, say, 55-- imply dissipation, that it had dropped all its moisture.  But it still needs watching.  It might not spawn tornadoes, but still could involve high winds and hail.  Bad cells tend to turn to the right but they can sometimes turn to the left. They tend to turn to the right since the indrafts come in from the lower right flank causing new cells to form in the lower left flank as the older cells dissipate.  The movement can be "apparent," though."

 

Needed

information

 

 

 

 

·        In some circumstances the needed data is spares and very limited. 

·        In some cases, direct observation (skywatching) is the best source of information.

·        Storm tracking either by the NEXRAD algorithm or by manual annotations is critical in the monitoring of severe weather.

 

"NEXRAD does storm tracks every 15 minutes.  You can also do it yourself by laying down marks. You can put marks on the radar screen.  It is a good aid.  You can put your own dots down with time marks.  We do this for hurricane eye walls.   I did it now and then to see which way the cell was going and see where it would go."

 

·        Needed information may be obtained only after long periods of data monitoring (e.g., NEXRAD)

·        Severe weather events are not always noticed or discussed in products from the NWS or the Weather Channel.  Over-reliance on those sources can lead to errors.

·        A change in the "V" pattern to the difluence aloft as seen in GOES imagery  shows that the difluence aloft is thinning out, implying that the cell is weakening.

 


 

Hypotheticals and Mental Modeling

 

 

 

·        Night shift affords the opportunity for more in-depth analysis of weather dynamics.

·        In some circumstances the needed data is spares and very limited. The forecaster must be able to engage in hypothetical reasoning and make assumptions about possibilities.

 

"It was a heavy cell, but too far from the Mobile radar to be able to see any tornado signature (for that you ideally need to be within 60 miles of the radar). I was concerned that it might spawn tornadoes."

 

·        A dissipation  or thinning out in the "V" pattern to the difluence aloft as seen in GOES imagery implies that a cell is weakening.

 

·        For supercells, when all the dynamics line up there is great potential for severe weather.  The forecaster needs to be able to reason hypothetically about the dynamics (i.e., features including a hard core in the lower quadrant, high tops, consistent movement, and storm tilt).

·        Storm dynamics involve "virtual turning."

 

"The turning you see in supercells is virtual.  It is not that the system turns, it is a matter of where cells dissipate and new ones form."

 

·        Storms and supercells can sometimes turn in anomalous directions.

 

"Systems will turn toward the left only if there are upper-level winds driving in, winds up between 500 and 700 millibars do the steering.  The last case I saw of one turning left was a storm moving east to west across the panhandle.  It turned left and moved over Pascagoula."

 

·        Maturing and dissipating storms have unique dynamics.

 

"The other dynamic is in older storms.  They divert the upper air and can push the newer storms.  One of them can hook around the other.   Sometimes they get caught up in each other's flow and merge and turn into really big storms."

 


 

Options

 

 

·        In some weather forecasting situations, even in severe weather, there are no alternative courses of action and no options.

·        Even if severe weather seems to be dissipating one still has to keep monitoring it though an option would be to stop monitoring. 

 

"If there had been a change in size--shrinking to 20 miles in areal extent or the tops getting less tall or reduced dBZ from 65 down to, say, 55--all this would imply that it was dissipating, that it had dropped all its moisture.  But you'd still watch it.  It might not spawn tornadoes, but still could involve high winds and hail.  Even a novice would keep watching it."

 

Goals

 

 

 

·        Always forecast for the worst possible situations (warnings, etc.). 

·        Be aware of the needs of the various clients and the circumstances in which they experience weather impacts (e.g., night- time supercell would not influence training flights but would impact Search and Rescue operations if they were needed).

 

Situation

Assessment

 

·        Novices may fail to notice small features that are "upstream" and not salient but that can develop into severe weather that impacts operations.

Time/effort

 

 

·        Severe weather situations involve a need to monitor data for prolonged periods of time (e.g., manual plotting of storm tracks over a period of many hours).

 


 

 

 Lessons Learned

 

 

Implications for Training

Weather situations can be standard scenarios in some respects (e.g., storm dynamics and development) but not standard in other respects (e.g., time of day, time of year.  Training on severe weather scenarios should focus on standard scenarios but also expose trainees to scenarios that depart in various ways from the standard scenarios.

 

"It is never easy.  You never have all the data. You have to make assumptions. That's where experience comes in. You know to look for these sorts of things."

 

"They had a young forecaster there, just out of Italy, new to the area and hadn't worked any severe weather. I turned him into a bird dog at just the right moment. He was within 2 - 3 hours of getting hit and he hadn't even been watching it."

 

"This was not a routine situation since it is not usual to get a heavyweight supercell at midshift.  Ninety percent of the time severe weather is rare for that season.  But this was a standard scenario in terms of the storm development and dynamics, but not standard in terms of the time of year and time of day of the storm.  Fast moving cold fronts coming from the west usually determine our winter weather--storms and small lines of storms.  A big cell developing in a south-west flow is rare for winter in the Gulf region."

 

 

Leverage Point  -  Re-design of METOC Operations floor

 

"I stood up and did a 360 looking at the Wall of Thunder and the NEXRAD."

 


 

 

Leverage Point -- Mechanism for information sharing and weather discussions among Gulf region facilities

 

"I assumed he got busy--doing his checklist--even before USAF Weather Center would have put out a Point warning, which would trigger his checklist. So I did not bother him after that. I never went back to see what he did.  I would not have called Tallahassee.  And Norfolk would have been no help either.  NWS Severe weather work is done at Kansas City.  We don't call them.  It would be good to be able to discuss sever events with them but they would never go for it.  If every Tom, Dick, and Harry called them it would pull them off their job.  It would be nice for there to be a checklist for NASP, Eglin and Hurlbert to do more information sharing and weather discussion, especially on the night shift (on day shift there's lots of data floating around)."

 

Concept Mapping Toolkit
Insitute for Human and Machine Cognition
The University of West Florida