The Westminster, TX Tornado of May 9th, 2006

On the late evening hours of May 9th, 2006, one of the most violent tornadoes in the WFO of the National Weather Service at Dallas Fort-Worth. The slow-moving tornado was on the ground for 9.3 miles, had a width of 300 yards, killed 3 people, and injured 10 people. This blog post will dive deep into the meteorology behind the tornado, an examination of the damage path the tornado left in its wake, and my opinion regarding the controversial rating of the tornado.

Meteorological Analysis

The Westminster tornado was triggered by a classic Colorado Surface Low. Several surface lows that have resulted in famous tornado events in the past have formed from lift created by the Colorado Rockies. Over the next day, the low slowly moved southeast and deepened (intensified) at a rather slow rate. Late on May 8th, a warm front pushed through the Dallas Fort Worth Metropolitan area, beginning destabilization ahead of the surface low. As the day progressed, the air continued to become more and more unstable, and more and more conducive for severe weather. However, a trigger was needed to result in supercell initiation, or else the environment would go unused. Lift is always triggered by a surface low, no matter the intensity of it. The surface low became the trigger for severe weather. The surface low also influenced favorable wind profiles for tornadoes, with surface winds out of the southeast.

(Above) Surface Analysis at 0000 UTC on May 10th, 2006. The location of Westminster is highlighted by the black dot.

Every day, usually at 1200 UTC and 0000 UTC, the National Weather Service releases a balloon into the atmosphere. This balloon is not your regular, average balloon, however. It is specifically engineered to collect data from the atmosphere, to help forecasters understand what the atmosphere is capable of. At 1200 UTC on May 9th, 2006 (approximately 7:00 AM CDT), the National Weather Service released a balloon to collect data from the atmosphere above Dallas. The environment had already destabilized quickly by this point, with the warm front having lifted through more than 12 hours earlier. Overall, the environment was quite favorable for tornadoes with the exception of one mesoscale aspect, the capping inversion. The capping inversion acts like a lid on the environment. It prevents hot air from rising up, preventing supercell development from occurring until it either weakens or is overcome by extreme levels of forcing along the cold front or by an unusually intense surface low. In this particular instance, the surface low was not that strong, which eliminated the possibility of overcoming the capping inversion using lift from the surface low alone. The trigger would have to come from a weakening of the capping inversion, or the cold front.

(Above): 1200 UTC Sounding Observations from NWS Dallas Fort Worth on May 9th, 2006

By 0000 UTC, the capping inversion had vanished into thin air. However, the wind shear values and hodograph had declined. The Critical Angle increased significantly increased. Those three mesoscale aspects had declined significantly for tornadoes, but there were also other mesoscale aspects of the environment that had significantly increased favorability for tornadoes. As the day continued, the environment continued to destabilize more and more. Surface heating caused the capping inversion to fade completely, favoring supercell development. The updraft profile had become so thick, vertical, and strong that if any supercell were to form, the updraft would be extremely efficient at turning the wind shear available into a tornado. Remember, the more vertical the updraft profile, the more vertical the mesocyclone, and the more efficient the mesocyclone becomes at using wind shear. Finally, the surface low was closing in on Dallas. The trigger for supercell thunderstorm developments had arrived. Everything was in place for a day of intense severe weather with some tornado potential.

(Above): 0000 UTC Sounding Observations from NWS Dallas Fort Worth on May 10th, 2006

Analysis of the Parent Supercell Thunderstorm

The supercell thunderstorm which produced the Westminster Tornado first developed at around 01:20 UTC, on the southern end of a line of severe thunderstorms. The updraft quickly became super cellular, and returns of >50dBZ began to come back to the radar by 01:30 UTC. The supercell initially moved northeastwards with the squall line, but suddenly deviated and began moving further east by 02:00 UTC. This eastwards deviation enabled the supercell to ingest more wind shear and SRH in the environment than those storms along the squall line to the north did. The supercell underwent a period of reorganization beginning shortly after 02:10 UTC, as the main northern cell began to be overtaken by a new cell to its south. The two had combined into one healthy updraft by 03:15 UTC. The Westminster Tornado began at 03:25 UTC.

Evolution of the Westminster Tornadic Supercell Thunderstorm on Radar

01:15 UTC

The supercell first appears on the southern end of the squall line

01:30 UTC

The supercell begins to return values of >50dbZ

01:45 UTC

The supercell begins to deviate eastwards from the main squall line, which continues northeast

02:30 UTC

The supercell undergoes a period of organization, as a new cell to the south overtakes the older one to the north

02:55 UTC

The supercell is nearly finished reorganizing

03:15 UTC

The supercell has completed the merger, and is beginning to show evidence of rotation. Tornadogenesis occurs 10 minutes later.

Analysis of the Parent Supercell Thunderstorm from NEXRAD Radar Data

The supercell thunderstorm that produced the tornado first began to show a hook echo on NEXRAD radar at 03:15 UTC. The circulation passed near Anna at 03:25 UTC, which is the likely time of tornado genesis. It is officially listed as 03:37 UTC, but this is likely incorrect. The hook echo gradually occluded as the supercell thunderstorm moved east, and the tornado eventually dissipated as a result of the mesocyclone occlusion. NEXRAD radar also allows us to look inside of the parent supercell, and to allow us to identify the nature of its updraft, where the strongest winds are, and where the strongest precipitation is located. A volumetric scan of the supercell, while the tornado was in progress, revealed the mesocyclone to be vertical. This means that the mesocyclone was extremely efficient at using the wind shear available in the environment, and this is in accord with the picture painted by the observed soundings. The supercell's anvil reached a maximum height of 59,400 feet according to the echo tops product.

One unusual aspect of the radar data available was how weak the mesocyclone appeared to be, for the intensity of the damage that was produced. The mesocyclone's maximum gate-to-gate shear value (a measurement of how intense the strongest winds in a mesocyclone are) only reached 92.1 mph, much weaker than other violent tornadoes in the past. However, the answer to this lies in the wind profiles of the environment. The parent supercell was located quite far from the radar. The radar beam was scanning the tornadic mesocyclone at a height of 5,997 feet. We'll just round to 6,000 in this case. 6,000 feet is equal to approximately 1.8 kilometers high. The wind direction at 1.8 kilometers high in the atmosphere, according to observed sounding data, was almost due west. This is identical, in fact almost exactly the same as the wind direction at 15 kilometers high. This means that there was very little directional shear in the upper levels of the atmosphere. The change in wind direction only became significant at the lowest levels of the atmosphere. Due to this fact, the mesocyclone was spinning much slower at a height of 6,000 feet than it was at 1,000 feet. It's very likely that had the Westminster Supercell occurred much closer to the radar, the rotational signature would have been much more impressive.

(Right): Observed wind profiles from the 0000 UTC Sounding Observations on May 9th, 2006

The Tornado's Damage Path

The tornado had a very slow movement forward. As a result, the damage path was not that straight, with several wobbles and changes in direction noted. The starting point listed by the NCDC is an area in a secluded field just east of Pepper Hill Drive, north of Highway 427. The starting point is about 3.48 miles northwest of Westminster. The tornado maintained a steady course to the north-northeast before it intensified as it crossed Highway 479. This is where the tornado's effects were first seen on satellite. The tornado rapidly intensified as it passed through an open field northeast of here. The tornado crossed over the Soil Conservation Service Site 13 Reservoir, producing very significant tree damage before it emerged onto another open field, just east of Highway 3133. The ground was scoured along a narrow swath in this area. Two homes were destroyed at the intersection of Highway 3133 and Highway 531. Fortunately, no fatalities were reported at this location.

(Above): Violent Tornado destruction at the Intersection of Highway 3133 and 531

More Damage from the Highway 3133 - Highway 531 Intersection

The ground scoured in an open grass field near the Intersection of Highways 3133 and 531

A home completely leveled to its foundation near the intersection of Highways 3133 and 531

Another home completely leveled at the Intersection of Highways 3133 and 531

The tornado reached its peak intensity as it moved north towards a subdivision of homes along Brangus Road. Two homes were swept away as the tornado approached the Brangus Road area. The tornado slightly weakened as the property of the Newsom Family suffered a direct hit. The tornado leveled their main home, and obliterated a mobile home nearby. Mary and Paul Newsom, an elderly couple, died when the tornado destroyed the mobile home. Their family members were injured. Several other outbuildings were damaged and destroyed in the vicinity. The tornado also produced very intense tree damage in this area. Aerial photos of the Brangus Road Damage showed a classic convergence damage pattern, with debris from destroyed outbuildings being wind-rowed into the main tornadic circulation. One more home was destroyed along Brangus Road before the tornado continued moving northeast. A Honda was thrown a long-distance and badly mangled around a tree. Its engine was ripped away.

More views of Damage from the Brangus Road - Intersection Area

An aerial of the Intersection of Highways 3133 and 531 and Brangus Road

A brick home that was completely swept away along Brangus Road

Another aerial of the Brangus Road area

The tornado then struck a home and an outbuilding along Colson's Hill Lane, a minor road just south of Black Road. Both were completely destroyed. A 14-year old boy was killed in the home when a chimney fell on top of the place where he was sheltering. The tornado continued to produce intense tree damage as it moved north, scouring the ground before striking an intersection between Yellow Bridge Road and another private road. Severe tree damage continued to occur as the tornado moved further to the northeast. The tornado began to weaken further as it approached a home along Gordon Road, which lost its roof and some of its exterior walls. All of the water in a swimming pool nearby was sucked out. The tornado weakened further as it moved northeast, destroying one more building and producing more tree damage as it crossed Highways 121 and 160 before dissipating.

Aerial Views of Damage for the final part of the damage path

An aerial view of the Colson's Hill Lane and Black Road area

An aerial view of the Damage along a Private Road near Yellow Bridge Road

An aerial view of a home with significant damage along Gordon Road

Map of the Tornado's Track

At Left is the track of the tornado that I have created, which is different than the official track the NWS had. Download an interactive KMZ file of it here (NOTE: You can only open it in Google Earth)

Rating Discussion

The rating of the Westminster Tornado has always been a topic of major discussion. The National Weather Service, despite the damage indicators and contextual evidence, rated the tornado as a high-end F3. They specifically noted in their survey that the slow-movement of the tornado may have increased the severity of the damage. However, recent evidence has shown this to be false. For example, the May 28, 2013, Bennington, Kansas EF3 Tornado moved very slowly for its entire lifespan, despite a wind speed of 247 miles an hour being recorded inside. It failed to produce any form of ground scouring. Several other famous and violent, slow-moving Plains Tornadoes have failed to produce ground scouring. However, the Westminster Tornado was able to. This evidence suggests that although the slow forward speed of a tornado may exacerbate the damage it produces to some extent, it doesn't have too much of an overall effect.

Several tornadoes from the period of 2003 - 2006 likely attained lower ratings than they should have. This fear of possible humiliation was incurred by the La Plata, Maryland Tornado of April 28th, 2002. Inexperienced surveyors had given the tornado an F5 rating, based on homes being swept away. However, upon expert analysis, it turned out that these homes were poorly built. One home was actually downgraded from F5 to F1 , as it had been moved off of its foundation. Several offices feared that something similar would happen to them, resulting in very conservative ratings. Some notable tornadoes that were rated lower than they likely would have today include the 2004 Harper, KS and Marion, ND Tornadoes, and the 2006 Gallatin Tornado. The Westminster Tornado was one of these. The next year, the Enhanced Fujita Scale would be released, and rating issues would be set in stone. I am confident that had this tornado occurred today, given the NWS stating that some of the demolished homes were "well-built," the scouring and debarking of trees nearby, and other extreme instances of contextual damage, the tornado would have received at least a high-end EF4 rating.

Comments