Chief Yeoman Mike Shea helps a crane operator move a wrecked vehicle during tornado clean-up efforts in Joplin, Missouri, June 2, 2011. A recent investigation by the National Institute of Standards and Technology yielded 47 key findings and 16 recommendations to bolster resiliency against such events. Wikimedia Commons/U.S. Navy photo by Lieutenant J.G. Ryan Sullivan
An investigation by NIST found that many factors contributed to the single deadliest tornado in official U.S. records and presents recommendations to bolster resiliency.
December 17, 2013—Shortly after a powerful tornado cut a swath of devastation up to 1 mi wide and roughly 22 mi long through the city of Joplin, Missouri, and the surrounding area, a team from the National Institute of Standards and Technology (NIST) arrived to conduct a field reconnaissance mission that soon led to a full-scale investigation.
“This was the single deadliest tornado on record since the U.S. began keeping official records in 1950,” says Marc L. Levitan. Ph.D., A.M.ASCE, the leader of NIST’s research and development program under the National Windstorm Impact Reduction Program (NWIRP), Materials and Structural Systems Division. “And one of the questions is why?” he asks. “What was different about this event?
“We found during our preliminary reconnaissance that the warning time was greater than the National Weather Service average warning time for tornadoes,” he continues. “The City of Joplin was very proactive in adopting the latest model building codes and enforcing [them], so there didn’t appear to be any problem there.”
The investigation’s goal was all-encompassing and straightforward: to determine the factors that contributed to the 161 deaths and more than 1,000 injuries in Joplin and the destruction of close to 8,000 homes and nonresidential buildings. Investigators looked at information about the tornado’s near-surface wind field, building performance, emergency communications, and public response to the threat.
“We found that there were many factors that contributed to the deaths and destruction, which ultimately led to a wide-ranging span of topics that were covered by our recommendations,” Levitan says.
NIST recently released a draft of its final report on the investigation, opening a public comment period that will conclude on January 6, 2014. The report presents 47 key findings from the investigation as well as a set of 16 recommendations to enhance the understanding of tornado hazards; improve building codes, standards and practices; and enhance emergency communication procedures.
The team conducted an exhaustive investigation of the May 22, 2011, tornado, examining factors in each of the fatalities, researching a variety of buildings that collapsed or were damaged, examining building plans for some of the structures to discern the wind forces that would have been required to create the observed damage, and examining tree-fall patterns to reconstruct the wind environment during the tornado.
The team was able to develop a detailed wind-field model of the tornado as it moved through the city. NIST estimates the maximum wind speeds were approximately 175 mph. Accounting for uncertainty, that estimate has a top range of 210 mph.
“In the final analysis, the performance of the buildings that we saw wasn’t surprising when you have an event with wind speeds that were so much greater than what we consider in our basic building codes and standards,” says Levitan. “Tornados are not a hazard that is addressed in our building codes and standards except if you are specifically designing a tornado shelter, a tornado safe room, or safety-related structures of nuclear facilities.”
The investigators found that of the 161 deaths, 155 resulted from blunt force trauma. More than 80 percent of those who died—135—were inside a building. The NIST report noted that there were no deaths in underground shelters and basements. All are uncommon in the area because of several factors, including a high water table and the presence of old lead and zinc mines beneath the city. There were no public shelters or safe rooms in the city of Joplin at the time of the tornado.
For the majority of the buildings that collapsed, the roof system was the key structural element, providing overall lateral stability, the report notes. When strong wind-induced uplift caused a failure of the roof system, the lack of support at the perimeter walls caused them to collapse under lateral wind loads. This was validated by structural calculation, interviews with eyewitnesses, and in one instance, a video recording from a security camera. Structures with redundant lateral load systems, or with a rigid concrete roof deck, withstood the storm without structural collapse.
NIST recommends that ASCE and other stakeholders in the industry develop “nationally accepted performance-based standards for tornado-resistant design for buildings and infrastructure.” NIST also recommends that model building codes include tornado shelters for both new and existing buildings and that shelters be installed at schools, mercantile buildings, structures in which large numbers of people are likely to gather, and multifamily residential buildings.
The report notes that measurements of tornado wind speeds near the ground are lacking, hampering efforts to create such national building codes and design standards. NIST recommends that the National Oceanic and Atmospheric Agency (NOAA), working in conjunction with academia, industry, and other stakeholder organizations, develop and deploy technology to measure and characterize near-surface tornadic winds and create easily accessed databases with this information.
The report also notes that confusion and distrust of emergency communication before the tornado may have played a role in the deaths and injuries. There was additional confusion because the tornado siren had already sounded a brief time earlier, on May 22, for a storm on the outskirts of the city that was moving away.
“As people were trying to get more information, one of the things that we found was the information they were getting was sometimes conflicting,” Levitan says. “Before the tornado was on the ground and the second alarm sounded, people couldn’t witness any cues. The first set of sirens were going off, but they didn’t see anything, they didn’t hear anything.”
“One of our key findings was that most people were seeking information and [when] they got some high-intensity cue, that’s really when they took action,” Levitan says. Examples include a television anchor imploring people to seek shelter immediately, or people seeing and hearing evidence of the tornado for themselves.
The NIST report recommends the development of national codes and standards for the dissemination of accurate, consistent emergency alerts and warnings. The report also recommends the full utilization of such “push” technologies as GPS-based mobile alerts and a sort of reverse-9-1-1 system that automatically calls residents to warn them of impending danger. Further, NIST recommends research into methods to enhance the public’s perception of personal risk in such emergencies.
At the close of the public comment period, NIST will review the information it receives on the report and incorporate any changes that will benefit the final report, which will be published in spring 2014. At about the same time, NIST will make available to the public the repository of maps and data collected and developed during the investigation.
“Hopefully this will be a good tool for other researchers and perhaps even practitioners who, ...maybe they haven’t seen directly the impacts [and] engineering analysis of what happens during a tornado,” Levitan says.
“Then we really roll up our sleeves, and we get to work. We have 16 pretty hard-hitting recommendations in this report. We need to work with the professional societies, we need to work with standards and code organizations, and we need to work with other federal agencies and other organizations in the profession at large to push as hard as we can to get these recommendations implemented. That will be a key focus and a key effort for NIST,” Levitan says.