A Case Study in Weather Forecasting:

Severe Weather in the Gulf of Mexico



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.

 

Although this case involves a typical scenario of severe weather developing over the Gulf of Mexico, the forecaster still needed to reason speculatively about what might cause intensification or dissipation of a storm system. Reasoning about how to determine the valid interval for warnings depends on thorough understanding of client needs and the activities in which clients engage as a result of issued warnings. Any need to ammend a warning implies a lack of understanding of the weather situation. It is important to have a thorough knowledge of typical scenarios, but also to have had enough lived experience so as to have had the chance to learn from errors during the typical scenarios.

 

 

 

 


 

 

Event and Comments

Event

Type

 

 

Time

Before arrival at METOC I was skywatching.

I saw cirrus to the southwest--anvil cirrus blowing off the tops.

You can see this even though the main clouds might be 100 - 200 miles away. It was not a blue or gold sunset.

A novice may or may not skywatch.

He may not have thought about the situation before coming into it.

Novices know they have others who will tip them off.

In this case, though, that would imply a delay before we would catch it.

 

Observation or situation assessment

March

1999


There were clear skies and high pressure over the region.

There were two Lows between here and Corpus Christi, TX.

It was like a stalled front.

The Lows were waves on a stationary front.

They were out over the Gulf.

One was SSW of New Orleans, the other was NNE of Corpus Christi, also over the Gulf.

The Lows were losing 1 millibar per hour.

This was a textbook scenario--a stalled front with waves and energy approaching it.

 

Observation or situation assessment

5:00 PM

Start of

Mid-

watch

This implied deepening.

There were no alternative interpretations.

This was open and shut.

All of the indicators were there--cirrus, lightning.

As soon as I saw the bouy data, that was it.

From the shift before we had a good analysis going.

The turnover briefing was short and to the point.

In the old days you'd do an analysis at the start of your shift and then look upstream at what will happen.

Now, we have so many monitors--all of the centers give us guidance or the big picture.

At watch change one shift does not have to say much to the new shift.

An experienced guy will now within 30 minutes what's going on.

But good pass-downs are useful.

The guy I was relieving was new.

He just gave the standard pass-down: Nothing was going on, a stalled front.

He had no idea of the potential.

At this point, I could have verified the strengthening and approach toward the Charlie Areas of Responsibility and started informing the customers.

 

Decision

 

At sunset I saw cirrus to the southwest, with lightning.

I could see anvil cirrus blowing off the tops.

I knew I had to cover Charlie 1 and alert to possible gale-force winds.

To a relief Officer at this point, I'd say that there were not enough data yet.

We had to query the bouys.

 

Observation or situation assessment

 

 

 

The following two Figures show the visible GOES images.

This implied that something in the atmosphere was turning over.

Southwest is the magic direction.

There must be some creature out there generating it.

Something strong was out in the "Charlie 1" area.

A novice would have known what to do if he remembered the scenario they are taught in the School.

No errors were likely at this point.

Something just had to be popping up the moisture and causing lifting and clouds.

 

Decision

 

I started analyzing the bouy data.

Winds were associated with the Lows.

I could see the pressure falling and knew I could put out a warning for strong winds.

LPATS showed a ring of lightning around the Low.

This was unusually symmetrical, but showed that the Low was well-organized.

From a hand plot of bouy data I could plot the front, the Lows (position, movement, rate of movement).

I did about one plot per hour, about 6 or 8 in all.

Enough to know that the warning had to go out and then 2 or 3 more plots to show that it really was out there.

After that I just checked the bouy data and added notes onto the plots I had done.

The analysis process is cut-and-dried.

You can see the trends in the pressure and wave heights.

If no bouy data were available, then this would be a tough call.

I'd look at LPATS and NEXRAD for the Doppler effect.

GOES for the cirrus.

But you can't get surface winds from NEXRAD because of earth's curvature below the horizon.

But you could read the winds a few thousand feet above the surface.

You might be able to get some ship reports--scan for them.

Also, oil rigs put out data.

You'd have to look for alternative sources of the surface data.

 

Action

6:00

PM

 

The following Figure illustrates the raw bouy data as received at the METOC, and the Figure after that is the original chart that Mr. Etheridge prepared.

 



I knew what the situation was.

There were no alternative interpretations.

My goal was to get ready to begin alerting our customers.

This was a textbook case.

A stalled front off the Texas coast.

You look out to the southwest and if you see any approaching trough, vorticity, or a vorticity maximum, any Low or wave on the front will develop one or two storm systems.

It is taught in the School and is discussed in the Local Handbook.

But you still need to experience it first-hand a few times.

Experience makes all the difference.

If you get burned once, then you learn.

Trainees are given training with sets of cases so they get exposed to it.

 

Decision

 

 

 

The following Figure shows the numerical model output indicating two regions of vorticity in the Gulf, indicated by "Xs" in enclosed dashed-line ellipses, one near the coast of Texas and the other south of New Orleans.

Pressures continued to drop in the Lows.

The forecast out of Norfolk had nothing.

The forecast out of Norfolk fit the scenario, actually.

They mostly focus on the Atlantic; they look east.

It was not totally unexpected that they would not be forecasting what I saw.

The Lows developing along the front could have blind-sighted Norfolk.

I wanted to alert them.

Norfolk should have put out guidance about something developing but little fronts in the Gulf are not their main concern.

If you are the little guy in the Gulf and a sleeper jumps up and bites you...

And they just might have been looking at it and working on it.

I needed to see what stage of the game they were in, so that we could avoid putting out conflicting warnings.

The pressures might have started going up again of the system causing the deepening went out of phase and the trough moved quickly.

Upper-air data covers only 12 hours.

Satellite fixes can be used to determine trough speeds, and you can look to see if the trough speed is greater than the speed of the Low at the surface.

If I lost the bouy data and the pressure started to decrease, I would have contacted other stations and have them give me the bouy data.

You're never really blinded.

 

Observation or situation assessment

 

I decided to call the Duty Officer at Norfolk.

There were no alternatives.

I went to them for confirmation since they are responsible for Gulf Warnings.

I asked: "Are you forecasting storms for here in Pensacola? Do you want to modify your forecast?"

The Officer replied that he had just gotten on duty.

 

Action

7:00

PM

He told me to cover my own area..

He looked at the data and agreed with me.

This confirmed my assessment.

And I knew that the customers would not get conflicting data.

 

Decision

 


 

I put out a non-tropical gale warning (calling for 34-47 knot winds), for Charlie 1 and Charlie 2, valid through the following day (18 hours out).

The warning could have been more specific (e.g., speeds).

It is better to over-warn than under-warn since you can always stop a warning.

But extending or changing a warning implies that o do not really know the situation

And when a warning goes out people go to work doing things like buttoning stuff down.

They may not be there to see a modification or an extension.

A novice would have done the same.

The NWS puts out bulletins at standard times for 12 -24 hour or long-range warnings.

But I had no time to look at those.

For fast-firing systems like this one (12 hours or less), the NWS bulletins do not help.

 

Action

8:00

PM

 

The following figure shows the Warning that was issued.

 

>

The following two visible GOES images show what appeared over the Gulf at 10:41 PM Central time and 2:12 AM Central time.



 

 

I called the Operations Officer, in case there were Navy operations going on.

People might have had special needs.

Calling the Ops Officer is checklist stuff you have to do.

 

Action

 

The trough moved across the Low and moved ahead of it and out of phase.

Subsidence caught the Low and weakened.

But at sea they had lightning (on LPATS) and gale-force winds (from bouy data).

It was unlikely that things would have unfolded any differently.

Maintained gale-force winds require major storm systems.

This is rare.

Storm of March 1993 hit western Florida with 112 mph winds.

that situation was similar to this one--everything lines up perfectly.

But major storms out of this scenario are rare.

These were minor storms.

 

Observation or situation assessment

Mid-

night

I kept monitoring the data--radar, LPATS, but the main data were from the bouys since they told winds and sea heights.

At this point, I was asking myself:

Were the Lows intensifying and moving eastward?

Intensification would imply a need to upgrade the warning.

Would people need to do preparations at the Base?

 

Observation or situation assessment

 

The Lows carried on for a while out at sea.

Based on the bouy data, I could see that the winds were flattening out and would dissipate.

The seas flattened.

The energy causing it had gone away.

Re-intensification was not possible.

There was only one energy source--the one trough.

 

Observation or situation assessment

 

Through the night.

Once the trough moves out of phase with the Lows, the pressure rose, the thunderstorms dissipated (seen in GOES), the sea heights dropped (seen in the bouy data).

Dissipation also fit the standard scenario.

My goal at this point had to do with the warning that had been issued.

The warning was for too long a valid interval.

Relief would have to change the warning to free people up.

There was no need for them to react to the warning.

They could go back to normal operations.

 

observation or situation assessment

 

I decided that the intensity was lessening.

Error was unlikely at this point.

A novice might have made an error if they did not keep looking at the bouy data at this point.

 

Decision

 

I explained the situation to the relief personnel at watch turnover.

They would have to decide whether or not to let the warning ride or cut it short.

I knew the warning would have to be cut short.

Under a gale warning, the watch change briefing would never be minimal.

The Operations Officer gives it a high level of attention since it has implications for the day's training activities.

If there were a need to extend the warning they would have to decide that at the beginning of the watch, during the turn-over.

If the warning remained in effect that would imply limits to what could be done in the Charlie areas.

 

 

Action

5:00

AM

Watch

change

They cancelled the warning.

 

Action

8:00

AM


 

Analysis

 

Non-tropical weather can kick off quickly from the warm water.

You need to look at where the stationary fronts are with weak Lows rippling along them and where upper-air troughs come across them.

The upper-air troughs move into phase with surface systems and for about 6 hours they intensify the Lows and make them pump up in their lower levels.

the trough causes difluence aloft and the air gets sucked up through the Lows.

This kicks off storms around the Lows.

Nor' Easters start the same way--a weak Low influenced by the upper-air.

Then the troughs overrun the Lows and fill them in.

 

 

Decision Requirements

 

Cues and

Variables

 

         Skywatching provides critical information before a watch period, as well as during it.

Needed

information

 

 

 

 

         A thorough knowledge of standard scenarios needs reinforcement by lived experience.

         Causes of lifting.

         Sometimes there is no effective substitute for surface data to inform about Low pressure systems, but helpful information can be gotten from LPATS (i.e., storm organization and intensification) and NEXRAD (mid- to upper-level winds).

         Understanding and forecasting of Gulf weather is critically reliant on data from bouys.

 

Hypotheticals

 

 

 

         Rules of thumb for hypothetical reasoning in standard scenarios can often be stated succinctly (e.g., stationary front over the Gulf with weak lows can be energized by upper-level troughs overrunning them from the south-west). "You look out to the southwest and if you see any approaching trough, vorticity, or a vorticity maximum, any Low or wave on the front will develop one or two storm systems."

         Forecasters need to be able to reason speculatively about what might cause intensification or dissipation of developing storm systems.

 

Options

 

 

         Mutual reliance among forecasters at various stations, to coordinate warnings and share information.

Goals

 

 

 

         Reasoning about determining the valid interval for warnings depends on thorough understanding of client needs and the activities in which clients engage as a result of issued warnings.

         Coordination of warnings among responsible forecasting offices.

 


 

Rationale

 

 

 

         Reasoning about determining the valid interval for warnings depends on thorough understanding of client needs and the activities in which clients engage as a result of issued warnings.

         It is better to extend a warning out for a longer rather than a shorter valid interval. Any need to ammend a warning implies a lack of understanding of the weather situation. Warnings can always be cut short at the watch change.

 

Situation

Assessment

 

         What might cause intensification or dissipation in a developing storm system?

         It is important to have a thorough knowledge of typical scenarios, but also to have had enough lived experience so as to have had the chance to learn from errors during the typical scenarios.

 

Time/effort

 

 

         In severe weather situations, hand chart-work and hand charting skills can be of critical importance to both understanding and forecasting.

         Monitoring of data for long periods (many hours) is sometimes necessary.

 

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