Facility and Community Tornado Shelters
Evaluating Existing Areas to Be Used as Shelters

Region VII
By Barie Brettmann, P.E., S.E.
Release Date: March 3, 2003

On the evening of May 3, 1999, tornadoes tore through parts of Oklahoma and Kansas, leveling entire neighborhoods and killing 49 people. On May 10, 1999, the Federal Emergency Management Agency (FEMA) deployed a Building Performance Assessment Team (BPAT) to assess the damage caused by the tornadoes. The BPAT’s investigation made it clear that a severe wind event can cause a large loss of life or a large number of injuries in high-occupancy buildings and in residential neighborhoods where people do not have access to shelters (FEMA 342 Midwest Tornadoes of May 3, 1999).

Studies have been done since the early 1970s to determine design parameters for shelters intended to provide protection from tornadoes. FEMA, in association with Texas Tech University, in 1998 published guidelines for small, in-residence safe rooms in homes (FEMA 320: Taking Shelter from the Storm: Building a Safe Room Inside Your House). Based upon the observations of the damage from the May 3, 1999, tornadoes, FEMA developed guidelines for larger facility and community shelters (FEMA 361: Design and Construction Guidance for Community Shelters). FEMA 361 provides performance guidelines for the design and construction of community shelters, methods to evaluate potential shelters in existing facilities, and benefit/cost analyses (including a CD-ROM with the B/C Model software).

Evaluating Existing Areas to Be Used as Shelters
Observations in the BPAT’s investigations of the May 3, 1999, tornadoes showed extensive damage to areas in facilities that were designated as tornado refuge areas. This included many refuge areas in schools. It was observed that many walls of hallways, which were designated refuge areas, collapsed onto where the children would have been if the tornadoes had occurred during school. It also noted that many areas that were the safest in these facilities were overlooked by the owners as places of refuge. Many items not recognized as potential hazards during tornadoes were shown to cause excessive damage. These items included non-tied down equipment or propane tanks, adjacent light poles and telephone poles, gravel from gravel-ballested roofs, and other small items that became missile-like projectiles.

Based upon these observations, FEMA analyzed the observations and developed checklist criteria for ranking safe refuge areas. The objective of the checklists is twofold:

• To identify structural and non-structural vulnerabilities to tornado events.
• To rank a group of facilities to determine which have the least structural resistance to high wind forces and are in greatest need of retrofitting.

Using this method, several areas of the facility are determined to be possible safe areas. By analyzing each of the potential safe areas, this analysis will help the owner to determine:

• If the safest parts of the facility are being used for the tornado shelter (or locate one if there isn’t already a designated shelter).
• Adequacy of the strength, safety and size of the shelter.
• Possible ways to make the shelter areas safer.
• Whether it is advisable to design and build a new shelter per the guidelines in FEMA 361.

Site Assessment Checklists (From FEMA 361)
The checklists are designed to walk the user through a step-by-step process and should be filled out in sequence. This process is a rapid visual screening and does not involve any destructive testing or detailed engineering calculations. A large portion of the checklists can be filled out using data obtained from design or construction plans. It is important to verify this data during a field inspection and note upgrades (i.e., expect roof replacements on older buildings). If building plans are not available for this evaluation, the accuracy of the checklists is compromised. Additional information can be acquired from building specifications, site visits, and interviews with building maintenance personnel who can provide historical information on specific problems, repairs, upgrades, and school procedures. The checklists consist of six parts.

1. General building information:
   This section is for collecting information for reference purposes. All questions relate to the entire building or buildings at the site. The user may need to refer back to the General Building Information section to answer hazard-related questions in other sections. This section is not scored.



2. Selecting the refuge area:
   The focus of the evaluation is to select appropriate refuge areas that might provide protection from high wind and tornadic events. The criteria contained in this section will guide the user on how to select good candidate refuge areas. Several refuge areas may be needed to provide enough usable space for the entire population in need of protection. A separate checklist should be filled out for each potential refuge area. This section is not scored.



3. Wind hazard checklist:
   This checklist applies only to the refuge areas. If more than one area is selected, a separate checklist should be filled out for each area. A glossary with diagrams is provided to help the user with unfamiliar terminology. The checklist questions are answered and a score is determined for this hazard.



4. Flood hazard checklist:
   This section applies to both the refuge area and to the entire building. A Flood Insurance Rate Map (FIRM) is required to answer most of the questions in this section. The checklist questions are answered and a score is determined for this hazard.



5. Structural seismic hazard checklist:
   The checklist for the seismic threat pertains to the entire building. A Seismic Activity Zone Map is provided to help assess the seismic threat. The checklist questions are answered and a score is determined for this hazard.



6. Summary score sheet:
   After answering and scoring all of the questions in the checklists, the summary score sheet is filled out. The score sheet is used to compile all of the scores for each refuge area associated with each site for comparison. The total scores will enable the user to rank each building and its potential as an adequate refuge area. Transfer checklist scores to the summary score sheet to include subscores from the wind section for each refuge area evaluated. The highest area total wind hazard score should be placed in the highest wind hazard score block. The total score is the sum of the highest wind hazard score, flood hazard score, and seismic hazard score. The total scores will reflect the expected performance ranking of the buildings when placed in order from lowest to highest score (i.e., least vulnerable to most vulnerable structure).
   
   Low scores on the checklists indicate structural features that provide some level of protection. Higher scores indicate that a refuge area is more vulnerable to wind damage. The lowest possible cumulative score for Zone 4 (region most vulnerable to tornado hazards) is 20, and a refuge area with this score would likely provide significant protection from a high wind event; however, it is very unlikely that any building, even one with an engineered storm shelter, would have this score. For example, a pilot study of 10 schools in Wichita, Kan., (located in Zone 4) resulted in scores ranging from 56 to 161.

Conclusion
Based upon damage, injuries and loss of life from recent tornadic events, many government agencies, school boards, hospitals, large manufacturers, office building owners, developers and others have been investigating better ways of identifying and constructing tornado shelters for their facilities. The FEMA checklist system will help in evaluating and establishing the safest areas in their facilities for tornado shelters.

Barie Brettmann, P.E., S.E., is an associate structural engineer with the aviation and architectural division of Burns & McDonnell, and has almost 20 years of experience specializing in the design of public and private buildings.

©2002 Burns and McDonnel, with permission. Any copying, redistribution, or retransmission of any of the contents without the express written consent of Burns and McDonnel is expressly prohibited.