A Case Study in Weather Forecasting:

Thunderstorms



This case is told by Mr. Howard Graham, senior civilian forecaster at the NASP METOC.  He has been a forecaster since 1958, including service at Guantanamo Bay, service with the "Hurricane Hunters," deployment on the USS Forrestal, and service as forecaster at Jacksonville FL, Memphis TN, Rota Spain, Charleston SC, and Norfolk VA.  He has logged nearly 55,500 hours at forecasting and forecasting-related tasks.



In the case described here, there was a question of where and when severe weather would errupt, but the forecaster disagreed with the warnings issued by National Weather Service.  When should a forecaster distrust the guidance, and how should he proceed in issuing his own warnings?  When should one deliberately over-forecast?

 

Event  and Comments

EVENT

TYPE

 

 

Time

(1).  At 5:00 PM

I began  Midwatch. Since it was the first day on midwatch, when I arrived I was not aware of the current overall weather situation. There was a watch change discussion, and I examined what was happening on  Radar and LPATS (lightning detection network).

 

Observation or situation assessment

5:00

(2).  A Thunderstorm Condition II had been issued for Meridian,

effective until 9:00 PM.  I did not know it I would have to issue a warning for severe weather.

 

Observation or situation assessment

5:00

(3). A line of thunderstorms was moving into Mississippi, heading east.

 

Observation or situation assessment

5:05 -

5:15

(4). That portion was dissipating, or weakening.  The line of thunderstorms moving across, with a few cells at the front that you could see on radar and LPAT.  T-II was in effect for Meridian. This was a familiar scenario, so I knew how to deal with it. Relief would have been told at this point that a line was approaching with severe to the south.  Relief would have been told to watch it carefully, and that Meridian would not take the brunt. A novice could have overestimated the severity and may not have had the fortitude to resist the temptation to inflate the forecast to cover their backs.  Experience gives an advantage--confidence.

 

Observation or situation assessment

5:00

(5). I judged that the weather would not be severe at Meridian. The line in the region approaching Meridian could have flared  up or intensified, so I watched it carefully.  I never felt uncomfortable with this situation, never re-assessed  my judgments.  My judgment formed over about a 15-minute period. There was no alternative course of action. There was  never a moment of doubt about the data or that I had missed something. I did look at the satellite imagery but for a line of storms like that and LPATS are the best tools.  Base reflectivity mostly but I also looked at echo tops to see if the clouds were building.  They weren't.  If they had built up I'd have changed  my judgment.

 

Decision

5:15

(6). I monitored radar and LPATS

 

Observation or situation assessment

 

5:15 -

7:00

 

(7). I issued Thunderstorm Condition II  for Meridian, effective from 7:00 PM - 3:00 AM.

 

Action

7:00


 

(8). NWS issued a severe weather warning. 

 

Observation or situation assessment

 

7:00

(9). Meridian calls: Do I want to upgrade for Meridian?  No, I did not. It did not  make sense to issue a severe.  Meridian would have to do stuff.  I do not do "just in case" forecasting.

 

Decision

7:05

(10). I continued monitoring radar and LPATS. 

 

Observation or situation assessment

 

7:15 -

8:30

(11). The  line of storms was not  broad.  If a novice did not pay attention to the  radar they could have missed a flare-up.  I watched every frame of the radar.  I've  learned from experience that you have to watch these things constantly.  A novice  might not have looked at what was available, e.g., echo tops.  If the echo tops  showed an increase in cloud top altitude then I would focus my attention on that  area and my thinking would have changed immediately, perhaps to go to a severe.  I would have done that in a heartbeat.

 

Observation or situation assessment

7:15 -

8:30

(12). This confirmed the forecast.

 

Decision

7:15 -

8:30

(13). I knew the line would hit Meridian within minutes.

 

Decision

7:15 -

8:30

(14).  The line approached Meridian. Radar showed that heavy activity was to the south and was moving in a way not a threat to Meridian (east, not north-east).I never felt he was missing it.  There was no doubt or concern.

 

Observation or situation assessment

9:00

(15). The other forecaster and I decided to issue Thunderstorm Condition I for

Meridian.

 

Decision

9:00

(16). I issued Thunderstorm Condition I for  Meridian.

 

Action

9:00


 

(17). I re-issued a Thunderstorm Condition II, effective from  9:00 PM until 8:00 AM for Meridian.  I could have ended the T-II at any time.  I extended it

until 8:00 AM in order to have it still in effect when the watch would change.  I wanted it to run to a point where I absolutely would not have to worry about it, i.e., if it would have to be extended again during his watch.  I also had to worry about New Orleans. By extending it, I could forget about it.

 

Action

9:00

(18). I asked Meridian to keep sending him observations. This was crunch time.  I extended the T-II to free myself.  I could cancel it later if I had to, but otherwise I did not have to worry about it. I knew the line would hit Meridian within about 45 minutes. The Meridian Observer  would be skywatching all the time.  If he saw a funnel cloud or if local radio said  they were experiencing severe weather, I would need to know immediately.

 

Action

9:00

(19). I continued monitoring radar and LPATS.  The  line approached  Meridian.  It

was not broad. 

 

Observation or situation assessment

9:30 -

10:45

 

(20).  This confirmed the forecast.  Mistakes were unlikely at this point. There was no need to upgrade for Meridian.  I had no second thoughts.

 

 

Decision

9:30 - 10:45

(21). I  queried Meridian since the line of storms was right at Meridian,

 

Action

About 9:45 or 10:00

(22). Meridian reported brief  and small  hail, winds  of about 27 knots. Nothing

could have been different.  I was confident that if the data were correct, I was

correct.  So I wanted information  from the Meridian Observer (e.g., was hail

occurring?).  My forecast was verifying.

 

Observation or situation assessment

About 9:45 or 10:00

(23).  This confirmed the  forecast.

 

Decision

10:00

(24).  The line passed through Meridian.

 

Observation or situation assessment

 

10:00

(25).  Thunderstorm Condition I expired, thunderstorm Condition II expired for Meridian.

 

Observation or situation assessment

10:00

(26). If severe storms hadn't occurred, they probably were not going to.

 

Decision

10:00


 

(27). Strong cell with tornado near Hattiesburg. I could see the cells on radar, and heard the reports on the Weather Channel.  Hattiesburg was getting pounded. Nothing could have been different, and the NWS was correct in issuing a severe warning.  The weather event was a done deal and everything was verifying. I do not recommend departure from NWS alerts but this was a clear case to me.  I'd have drawn the same box as the NWS did--to include the entire line of storms.  But I  had responsibility for Meridian. I take it case-by-case.

 

Observation or situation assessment

About10:00

(28). The line had moved to the east.

Observation or situation assessment

 

10:00 to

11:00

(29). The Meridian Observer and I agreed that Meridian was OK

 

Decision

10:00 to 11:00

 

(30). Meridian civilian airport ASOS reported maximum gusts of about 32 mph.

 

Observation or situation assessment

 

10:00 to 11:00

(31). This confirmed my forecast.

 

Decision

10:00 to 11:00

 

(32).  Radar showed that the line was weakening as it moved to the east, even the southern portion of the line. If the line of storms hadn't weakened I might have had to issue a T-II for NASP.  If it had strengthened I would have had to issue a severe. A novice might have issued a T-II because of being nervous about mis-judging its wind speed.  But I was still confident--felt no doubt.

 

Observation or situation assessment

11:00

(33). Thunderstorm Condition I issued for NASP and Whiting Field.  NASP would get rain and not much else.

 

Action

Midnight

(34).The line of storms had weakened and was not so severe at NASP.  Winds were about 20 knots and hail was about 1/2".

 

Observation or situation assessment

 

1:00 or 2:00


Decision Requirements

 

Cues and

Variables

 

 

 

 

·         Radar reflectivity (precipitation)

      Radar echo tops (cloud top altitude)

      LPATS (lightning network)

      GOES (satellite) imagery

                  as  indicators of:       storm activity,

                                                   storm intensity, and

                                                   storm strengthening/weakening

·         For frontal systems, radar base reflectivity and LPATS are the best tools.

·         Direct observation via skywatching in the region where severe weather is occurring, even at night.

·         Wind speed and direction.

·         Precipitation rate

·         Hail size and duration

 

Needed

information

 

 

 

 

·         Awareness of overall weather situation when beginning a watch.

·         Awareness of current warnings and alerts when beginning a watch.

·         Radar data, frame by frame.

·        Direction in which fronts, storm cells are moving, and the rate at  which they are moving.

Hypotheticals

 

 

 

·         It can be easy to miss a flare-up or strengthening in storm cells, unless one keeps examining the data.

·        Do not deliberately over-forecast for fear of mis-judging storm severity.

Options

 

 

 

 

·         Be sure to examine all the available and data pertinent to monitoring severe thunderstorms--radar echo tops, GOES imagery.

·         Try to reduce mental workload by extending the valid interval for a forecast so that attention can be distributed to other areas of concern.

·        Be willing to communicate with distant observers/forecasters to get timely information.

 

Goals

 

 

 

·         Careful inspection of data stream to watch for development vs. weakening of storm cells.

·         Avoid overestimating storm severity and resist the temptation to inflate the

       forecast to just to "cover your back."   

 

Rationale

 

 

 

·         Avoid deviating from NWS watches and warnings but be aware of clear cases

     when departure is warranted, e.g., an NWS warning box covers an entire line of

     storm cells but it is clear that the cells will be severe only in one region.  Be willing

     to take it case-by-case.

 

Situation

Assessment

 

·         It is critical to be familiar with the dynamics of frontal systems, and their manifestations in the local climate.

Time/effort

 

 

·         It is important to possess the willingness and fortitude to inspect incoming data stream frame by frame for long periods of time.

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