A possible tornado struck Troy, Alabama, without warning on Thursday night, damaging several structures and overturning a tractor trailer along its path. Early reports on Twitter indicate that the local Walmart and a sports store sustained heavy damage, with the roofs of both stores caved in, much to the surprise of the dazed shoppers inside.
Troy—home to 18,000 people and the 20,000-strong Troy University—is located in the southeastern part of Alabama, about halfway between Montgomery and Dothan. The Troy Messenger reported on Thursday night that there didn’t appear to be any serious injuries in the incident, which is incredible given the damage pictures rolling in.
Numerous pictures posted to Twitter in the hours after the incident show some damage to the exterior of Walmart and the adjacent Hibbett Sports, with the roofs of both stores caved in. Box stores are notorious for their inability to withstand even a weak tornado, as the structural support isn’t there to handle strong winds or debris.
The National Weather Service often does an incredible job when it comes to warning us of dangerous weather on the horizon, but for a variety of reasons, they can and do miss things on occasion. Tonight was one of those occasions.
The storm that passed over Troy late on Thursday night was a supercell with a pronounced hook echo and strong rotation on radar, which became obvious around 10:24 PM CDT:
You’re looking at two views of the same map in this radar snapshot. The left-hand image shows base reflectivity, or the common precipitation depiction we’re all used to seeing. The right-hand image shows base velocity, or the winds in the thunderstorm. Bright green shows fast winds moving southeast, while the adjacent reds show wind moving northwest. Where the two colors are tightly packed in what’s known as a couplet, it indicates strong rotation and a possible tornado. The location of the Walmart and Hibbett Sports is indicated by the small red dot southeast of Troy.
Here’s the radar sweep from 10:34 PM CDT—ten minutes after the rotation first became apparent on radar—or about the time that the possible tornado struck the stores, showing continued strong rotation with a pronounced hook echo:
It wasn’t until five minutes after the storm passed through the area—10:39 PM CDT—that the National Weather Service office in Birmingham, Alabama, issued the tornado warning.
BULLETIN - EAS ACTIVATION REQUESTED
NATIONAL WEATHER SERVICE BIRMINGHAM AL
1039 PM CDT THU AUG 6 2015
THE NATIONAL WEATHER SERVICE IN BIRMINGHAM HAS ISSUED A
* TORNADO WARNING FOR...
SOUTH CENTRAL PIKE COUNTY IN SOUTHEASTERN ALABAMA...
* UNTIL 1115 PM CDT
* AT 1039 PM CDT...A SEVERE THUNDERSTORM CAPABLE OF PRODUCING A
TORNADO WAS LOCATED OVER TROY UNIVERSITY...OR NEAR TROY...MOVING SOUTHEAST AT 20 MPH.
SOURCE...RADAR INDICATED ROTATION.
IMPACT...FLYING DEBRIS WILL BE DANGEROUS TO THOSE CAUGHT WITHOUT SHELTER. MOBILE HOMES WILL BE DAMAGED OR DESTROYED. DAMAGE TO ROOFS...WINDOWS AND VEHICLES WILL OCCUR. TREE DAMAGE IS LIKELY.
The office did issue a product known as a “significant weather advisory” more than an hour before the storm’s arrival, advising residents in the path of the storm that they could expect wind gusts up to 40 MPH, torrential rainfall, and frequent lightning. These products rarely reach the public, however, and there was no mention of a tornado or any potential for damage from the storm.
Usually when a National Weather Service office fails to issue a tornado warning on a storm, it’s because the tornado happened too fast—occurring between radar sweeps—or it formed below the radar beam, making it invisible to our fallible technology. Neither of those scenarios was the case this time. Furthermore, the area was under a marginal (one out of five) risk for severe thunderstorms tonight in the latest forecast from the Storm Prediction Center, but the agency did not have Troy under a risk for tornadoes.
NWS Birmingham covers one of the most tornado-prone jurisdictions in the United States—central Alabama is the unwilling host to violent tornado outbreaks from time to time, most infamously finding itself as ground zero for the disaster on April 27, 2011, which became the worst tornado outbreak ever recorded, producing 211 tornadoes from Mississippi to New York over the course of 24 hours.
The meteorologists at this office are experts at what they do, and they’ve led the agency in reducing the number of false alarm tornado warnings over the past couple of years. However, even the greatest expertise and biggest accomplishments don’t much matter to the folks who endured this possible tornado without warning. It took the office 15 minutes to issue a tornado warning on a storm that displayed obvious rotation, apparently coming too late for the people of Troy, and especially for those caught inside the Walmart that was torn apart tonight. There wasn’t even a severe thunderstorm warning in effect ahead of the storm, which makes this case even more unusual in the grand scheme of unwarned tornadoes.
It’s rare that people can legitimately claim that they had no warning before a tornado. This is one of the occasions where they can say that without stretching the truth. This cannot happen again, and the meteorologists at the office in Birmingham need to identify what went wrong so they can prevent a similar scenario from happening again one day.
I’ve reached out to NWS Birmingham for comment on why they did not issue a tornado warning sooner, and I will update this post when they respond.
NWS Responds (Update, 1:53 PM Friday)
Jim Stefkovich, the meteorologist in charge for NWS Birmingham, provided me with this response via email:
A significant weather advisory was issued around an hour before the tornado developed. At 1037 pm, the strongest rotation was shown on radar, a delta V of approximately 30 mph at the lowest elevation angle from the Ft. Rucker. WSR-88D. It was decided a warning was issued. As one person received a report in the office at 1038 pm, the warning operator completed the warning, which was transmitted within 30 seconds of the report at 1039 pm. Again, we will be looking at more data, but from all appearances, the rotation weakened in the very next volume scan.
Rotation like this occurs in numerous storms in AL, many of which produce no damage whatsoever. Couple this with the fact that the rotation was weak, and for only one volume scan, made this a very difficult warning challenge, and in my 30+ years of experience, when tornadoes go unwarned, it is for these types of events.
Quite a few meteorologists (employed by both the NWS and private companies) have challenged me on my interpretation of “strong rotation” as stated in this article. I had the benefit of hindsight, writing this article just two hours after a likely tornado struck a shopping center without warning. As several of these meteorologists have stated, this rotation doesn’t meet the definition of “strong” based on scientific guidelines. Looking back at the radar imagery, there was obvious rotation embedded within a hook echo moving into a populated area ten minutes before the likely tornado hit the shopping center. “Obvious” doesn’t equate to strong by the definition of the term, and what looks strong on radar to me in fact looked weak to them. However, I’m not responsible for issuing warnings, and what the experts think is what matters in an unwarned tornado situation like this.
After reading Mr. Stefkovich’s email and speaking informally with other NWS meteorologists, it seems there was a decision not to issue a warning based on the radar data before the likely tornado, and this decision appears to be in line with NWS criteria and training. Walking the line between issuing a false alarm on weak rotation or not warning on weak rotation that does indeed produce a tornado is something the agency needs to continue to improve. We’re lucky there were only a few injuries and nobody was killed, or I wouldn’t be the only person making a fuss over the lack of a warning.
The original text of this article remains in place, unchanged.
[Images: Gibson Ridge]