Have you ever looked at a radar image on television or the internet and thought, “What are those guys at the National Weather Service thinking?  That storm doesn’t look severe!" 

The radar image you are seeing is only a fraction of the data that National Weather Service meteorologists have available.  Radar data is only one part of the equation that leads to a severe weather warning decision. 

A warning meteorologist must have knowledge of the storm environment and how that environment is changing over time.  He or she must also know the conceptual models for severe storms, be able to apply those models to observational data, and know how to properly analyze and interpret data from the Doppler radar.

All of these factors play an important role in determining whether a severe weather warning is warranted or not, and whether an event is warned or missed.  There is no doubt that the WSR-88D (Doppler radar) is our primary tool for warning decision making, but it has limitations.  Let’s take a look at a recent event that I personally worked to illustrate this point.

Thursday, January 4, 2007, saw widespread showers and thunderstorms develop over Southeast Texas and Southwest Louisiana as an upper level storm system approached the area from the west.  Although the region was placed under a slight risk of severe weather by the Storm Prediction Center, a lack of atmospheric instability kept the vast majority of thunderstorms non-severe, despite the more than adequate wind shear for severe storms.

At around 1:20 pm, thunderstorms began to intensify over the coastal waters to the south of Pecan Island, which prompted the issuance of Special Marine Warnings.  About 20 minutes later, much of south-central Louisiana was placed in a Tornado Watch by the Storm Prediction Center, with damaging winds and isolated tornadoes mentioned as the primary threats.  Individual storms were moving to the northeast, while the entire area of showers and storms was spreading to the east.  The thunderstorms on this day were relatively shallow storms, with the highest reflectivity cores on radar only extending up to around 25,000 feet.  (By comparison, thunderstorms over the Central Plains can have reflectivity cores on radar that extend above 40,000 feet.)  This was indicative of weak instability in the atmosphere that would limit the strength of the thunderstorm updrafts and thus the hail potential.  Given this storm environment, I was primarily concerned with damaging winds and possibly isolated tornadoes.

As the entire area of showers and thunderstorms continued to march to the east, storms that were initially offshore began to spread inland into Vermilion Parish, while new cells were continually developing over the coastal waters.  The thunderstorms spreading inland exhibited broad rotation on radar at the lowest elevation angle (0.5°).  The radar beam at this range across Vermilion Parish is approximately 5,000 feet above the ground.  At this height, the radar beam is penetrating the storm well above the region where tornado development occurs.  This is one of the unfortunate limitations of Doppler radar.  Because less radar data is available about the lower levels of a thunderstorm the further it is from the radar, warning decisions become much more difficult (especially tornado warnings).  In addition, the data resolution becomes coarser, which can result in some loss of data (especially velocity data) due to the averaging and processing done by the radar.

As the thunderstorms moved to the northeast and began to head farther inland into Lafayette Parish, their rotation improved slightly, but still remained below generally accepted guidelines for issuing a Tornado Warning.  Meanwhile, another storm with similar weak rotational velocity was crossing Marsh Island into Vermilion Bay.  This storm differed from the rest in that the reflectivity signature more closely resembled the conceptual model of a low-topped supercell, and this by itself suggested rotation.  Still, at this distance from the radar, the lowest elevation slice of the radar beam is now 7,000 to 8,000 feet above the ground. 

Watching these two storms concurrently, my immediate concern was for the storm over Lafayette Parish since it was nearing major population areas.  Once the rotational velocity signature associated with this storm began to weaken, my focus shifted immediately to the second storm, which was now nearing the Iberia Parish coast at around 3:20 pm.  Unlike the storm over Lafayette Parish, the velocity and reflectivity signatures were becoming more indicative of a potential tornado. As the storm moved inland into southern

Iberia Parish, the rotational velocity strengthened a little bit more, but still remained below tornado warning guidance (see above radar image).  Then, a final piece of the puzzle came when the weather observation from the New Iberia Airport showed a 40 knot wind gust (46 mph) from the south-southwest.  My assessment of the situation at the time indicated this was evidence of a forward flank downdraft, the presence of which has been shown to correlate with tornado development.  At 3:40 pm, I issued a Tornado Warning for Iberia Parish. With warning decision making being passed off to the next meteorologists on duty, I left the office at 4:00 pm not knowing whether my warning decision had been correct.  After arriving home, I received a call from the office informing me that a tornado had indeed touched down, with damage, injuries, and two fatalities.  As this situation shows, severe weather warning decision making can be very difficult at times, since on occasion thunderstorms that barely appear severe on radar can cause considerable destruction