BY CHRIS WESSELS AND STEVEN P. WOODWORTH
On March 14, 2008, at approximately 2145 hours, downtown Atlanta, Georgia, was hit by a storm that included an EF 2 (Enhanced Fujita scale) tornado. The storm touched down and left a path of destruction approximately six miles long and several blocks wide from the Vine City area, through downtown, and into the east Atlanta neighborhood of Cabbagetown. It was apparent early on that Atlanta had suffered a significant event with widespread damage that would put the Atlanta (GA) Fire Rescue Department (AFRD) to the test.
Moments after the storm ended, AFRD units responded to the Fulton Cotton Mill Lofts to a report of buildings with heavy storm damage. Initial arriving companies found that five of the seven buildings in the complex suffered damage; one building (Building E) had suffered two major collapses. The extent of damage triggered the response of AFRD’s Special Operations Command (SOC) and the mutual-aid response of five other Metro Atlanta technical rescue teams, which together comprise the Metro Atlanta Georgia Search & Rescue Strike Team (GSAR) (photos 1, 2).
 Photos by Greg Simpson. |
 2. |
THE STRUCTURE
The century-old Fulton Bag & Cotton Mills complex was once a thriving textile village southeast of downtown. Now called Fulton Cotton Mill Lofts, it is believed to be the largest loft conversion in the country. At this site, Atlanta Firefighter Matt Moseley made a memorable helicopter rescue to save trapped crane operator Iver Simms in 1999. There are seven buildings on the 12.5-acre site, most of them built between 1885 and 1922. Six of the structures are heavy timber construction (Type IV) and comprise 750,000 square feet divided into 506 units. All of the buildings within the complex had been occupied to some extent during the storm, and all had received some damage.
Building E, built of heavy timber construction between 1915 and 1918, sustained the most significant damage. The building has load-bearing masonry walls (brick) with heavy post-and-beam floor systems. The columns are round 16-inch-diameter posts that are notched for the beam connections. At each connection point, steel collars support the beams and two lags attached from the collar to the beams. The floor and roof beams are heavy 8- × 14-inch timbers; the floor and roof are 4- × 6-inch tongue-and-groove planking. The building was originally designed with a load capacity of 400 pounds per square foot. As part of the building’s renovation into lofts, a three-inch layer of lightweight concrete was added to the floors to provide a two-hour fire rating between floors.
The areas of concern included the fourth floor, which had received major structural damage throughout the entire floor area. Approximately half of the roof was gone. In addition, two different areas in the building had suffered pancake collapses from the roof through all four floors into the basement. The first area of pancake collapse was in the northwest corner of the structure (photo 3). The second area of pancake collapse was on the south side of the building, in the middle of the structure. The south side collapse had also severed the main water line entering the building. Water was impinging on the west wall, eroding the mortar from the bricks, and flooding the basement.
 3. |
On the building’s southwest end, an elevator was covered by collapse material from the fourth through the first floors; responders had received three reports of occupants trapped in the elevator.
THE RESPONSE
Initial-arriving companies had self-deployed within the complex and initiated primary searches. On arrival of SOC, the first priority was to determine the extent of damage to each structure and the corresponding stability and to determine in which buildings searches had been initiated and which areas still remained to be searched. All companies were ordered out of the structures. This allowed the officers time to brief the chiefs as to the extent of the operations conducted to this point while simultaneously allowing the command teams to gain accountability of all personnel on the scene.
Because of the large volume of emergency calls in the city, the initial arriving battalion chief was redeployed to another call. Command was passed to the SOC battalion chief, who had arrived on-scene. The captain of the squad was assigned to the Operations group until additional chief officers could be brought on-scene. This would allow the command system to remain functional; but as with most of the city, it was stretched thin. The incident commander (IC) requested the first two arriving Georgia Search & Rescue units be sent to the Fulton Cotton Mill Lofts complex. He also requested three additional chief officers, because of the size and complexity of the incident and to reduce the span of control for the large number of rescuers that would be needed to bring the incident to a conclusion.
With the arrival of additional resources (an off-duty AFRD assistant chief, a mutual-aid battalion chief, and a mutual-aid special operations chief), an incident command system with the proper depth was established. The AFRD assistant chief assumed the role of IC, the AFRD SOC chief headed Operations, the mutual-aid special operations chief headed Logistics, and the mutual-aid battalion chief served as the incident safety officer.
SIZE-UP
Exterior. It was determined early in the incident that three areas had enough debris so that an exterior search was critical. In these areas, the roof was constructed of 8- × 14-inch beams with 4- × 6-inch tongue-and-groove planks. Roof sections, some as large as 15 × 25 feet, had been blown off the building. The debris was in areas of the complex that were heavily traveled by the residents. The plan was to use an engine and truck company to breach and search these debris areas, using chain saws and forcible entry tools. No victims were found in this area.
Electric, gas, and water feeds served each building; it was not known to what extent the utilities in each structure had been damaged. Although the electricity was out in the area, as a precautionary measure, the main breakers in each building were turned off as well. Gas service to each building was also shut down in the complex, except for Building A. Water service was cut off to the entire complex because of the water main break in Building E.
Interior.The interior search began with an assessment of the structure’s stability. The connection points of the columns and beams were checked on the first, second, and third floors. Two crews were sent to conduct air monitoring of the entire structure during the interior size-up. Although the potential for gas leaks was great, no hazard was detected. Simultaneously, the structure was evaluated for stability. Although some damage was noticed throughout the structure, it was determined that the majority of damage on these floors was to partition walls; the load-bearing members were still in good condition. A transit was set up on the main area of concern on the south side of the structure. One member was assigned with a radio to maintain a watch over this portion and report any movement to command.
Command felt that because of the type of construction, despite the heavy damage, the building was safe for crews to conduct search operations. The member assigned to the transit did report some movement of the area of concern early on in the incident; however, this was determined to be the result of some settling of debris and posed little potential for secondary collapse. The fourth floor had sustained structural damage, and a large portion of the roof was missing. The most significant factor affecting stability was the water flooding the basement and impinging on the west wall. Command was able to control this hazard early in the incident by shutting off the water to the entire complex.
Aerial observation. Because of the size of the structure and the difficulty of operating during nightfall (with no ambient light because of the loss of power), it was determined that an aerial reconnaissance of Building E and the rest of the complex would be valuable in determining search methods and hazards.
Another AFRD special operations battalion chief, called in to establish the citywide incident management team, was selected as the aerial observer. There was a long enough break in the weather to allow him to conduct the aerial recon using one of the Atlanta Police Department’s helicopters. Using the helicopter-mounted searchlight, the SOC chief obtained an excellent view of the structure from above and provided a great deal of helpful information for planning search priorities. This information helped to identify hazards created by the collapse and to determine the stability of a dislodged roof section. This overall view of the incident would prove very beneficial as the Metro Atlanta All Hazards Incident Management Team began establishing an Incident Action Plan for the next operational period. The team now had a first-hand perspective of the entire incident. All personnel were removed from Building E during the aerial recon in case rotor wash caused any problems.
SEARCH PLAN
According to the plan, the IC would coordinate exterior searches and the Operations chief would coordinate the interior searches, allowing each chief officer to maintain an effective span of control. Whenever possible, the AFRD policy is to maintain a small span of control during technical incidents.
The resident manager provided a list of occupied and vacant units within the structure. In the interior search’s first phase, on-scene Atlanta crews would search the occupied units and the intact elevators. Once the first phase was completed and additional resources arrived, the second phase would focus on searching the pancake collapse areas and the elevator covered by collapse debris.
Eighteen units were deployed within Building E, broken down into geographic groups of six. Each company on-scene searched one group of units and the elevators within their assigned area. While searching the fourth-floor units, one firefighter received a significant injury to the left ankle. The company officer notified Command, and the member was removed to a waiting ambulance. Once the injured firefighter was removed and the void searches were completed, the progress and extent of damage were reported to and tracked by Command.
The search of each of the occupied units proved negative, much to the surprise of all involved. The elevator that had collapsed from the weight of the debris and the force of the storm would take a little longer. Residents of the complex claimed to have seen four people in the elevator as the storm was passing. Debris had collapsed on top of the elevator and caved in the roof. Rescuers forced open the doors using a hydraulic rescue tool and shored the elevator in place using pneumatic shores. Once the elevator was secured, rescuers entered the car and removed enough debris to ensure that the car contained no victimsonce again, much to the surprise of all involved. The elevator shaft was searched using video search cameras. This allowed members to gain a 360-degree color view of the shaft above and below the elevator car. This area also proved to be all clear, and there was no structural damage to the shaft.
The first search phase concluded with no victims located. The second phase could now begin, as more technical rescue teams arrived on-scene. Crews were now assigned to begin searching the two areas of the building that had suffered the pancake collapse. Interior walls were breached to allow the search crews access and to maintain egress routes. Command diligently maintained egress routes, since storms were still passing through the area. Once the void searches were finished and no victims were found, members conducted a technical search using listening devices. Command called for an “All Quiet” on the scene to ensure the accuracy of the search. Once again, to everyone’s surprise, the searches proved negative. Command had requested a dog team to search the structures earlier in the incident. On arrival, the dog search teams searched each area of the collapse while the other search personnel went to rehab. Again, the search proved negative.
Command had been able to account for all of the residents except one. The areas had all been searched by physical void search, by technical search, and by dog teams. Now the question confronting command was, Could the victim have been missed during the searches, or was there a search area that had not been accounted for? As crews continued to rehab, the chiefs met and reviewed the plans of the complex and Building E. They discussed the possibility that each search could have missed a victim. The crews had been operating for more than 20 hours; fatigue could surely be playing a factor.
After a short discussion, each of the chiefs agreed that first and foremost they had confidence in the crews, with whom they had trained and whose capabilities they knew.
Second, the odds that three separate searches had missed a live victim seemed unlikely. It was believed that the safest option at that point was to postpone any efforts until daylight, when the scene could be evaluated under better conditions. Crews were allowed to take up any equipment not deemed necessary to the operation and return to quarters to swap out with relief crews. The last occupant of Building E was accounted for later that morning, when it was learned that he had not been in town during the storm.
COORDINATION WITH THE CONTRACTORS
Quite early in the incident, the owner of the real estate development company and the project manager for the contractor arrived on the scene. They provided us with invaluable resources and information as we developed search plans and accounted for residents.
The developer provided a site plan complete with building identification designations as well as complete rosters of complex residents. In addition, he brought in staff to set up a temporary staging location for the complex occupants and worked to account for residents in the damaged areas of the complex. This effort alone greatly enhanced the search teams’ efficiency. The IC ensured that information flowed steadily between the command post and the developer as a link between the fire department and the residents.
Throughout the incident, there was a coordinated effort between the IC and the project manager for the site. The two biggest assets the project manager provided were a detailed construction drawing of Building E and insight into details regarding the construction of the building. The construction drawings were used to chart collapse areas and plan the search strategy for the building. The mill-to-loft conversion has been a phased project since the late 1990s, so the project manager had a detailed knowledge not only of Building E, which was the most severely damaged, but of the entire complex.
The fire main for the complex was configured in a loop from when the complex was an operating mill. Debris from Building E crushed the fire main loop at that building, which caused flooding in the building and necessitated cutting off water for the entire complex. With no power for the fire alarm system and no water for the fire suppression system, even the undamaged areas of the complex could not be occupied until life safety systems could be restored. AFRD worked in conjunction with the developer to provide fire watch while residents retrieved belongings.
LESSONS LEARNED
We learned valuable lessons from the Fulton Cotton Mill Lofts operations as well as others throughout the city. Although Atlanta has suffered numerous incidents of trees falling on houses and damage from ice storms, the March 14 tornado inflicted the most widespread storm damage in the city’s history. The fact that the storm occurred after darkness further compounded issues for the Atlanta Fire Rescue Department.
The Fulton Cotton Mill Lofts. Although we used portable lights powered by fire department generators during the operation, visibility was severely limited, hampering personnel’s ability to fully evaluate building hazards. In daylight, several hanging and suspended debris dangers were seen that were not fully recognized during the night operations. When conducting searches in limited-visibility conditions, command should focus on maximizing search effectiveness and limiting rescuers’ exposure. In addition, overall damage to the complex was more severe when seen in daylight (photo 4).
 4. |
The ability to cross-reference resident rosters and confirm residents’ whereabouts streamlined the search efforts. In addition, although we requested and used search dogs from DeKalb County (GA) Fire Rescue, they should have been requested earlier in the incident. The dog teams were requested approximately two hours after the storm had hit. The backlog of communications combined with an overload of the communications system caused a delay of several hours before the dog teams arrived. The search dogs are one of the most effective search methods when there is a need to complete a large area quickly.
Citywide operations. The unified command post was downgraded too early in the incident, which made coordinating efforts such as street closures and high-rise glass management more difficult. Although fire companies and collapse rescue teams searched damaged areas throughout the city where applicable, this should have been coordinated with damaged building assessments and hazard evaluations. Engine and truck companies searched numerous buildings prior to a thorough hazard assessment.
Coordination between the fire department and the building department expedited the efforts to get dangerous buildings slated for demolition.
High-rise issues.The storm damaged hundreds of windows in approximately 10 high-rise buildings downtown. Initially, although streets surrounding these buildings were closed, there was still a significant amount of pedestrian traffic in these areas. The AFRD conducted a thorough survey of the damaged high-rise buildings to determine the risk to the public.
Once the list of buildings with hazardous glass was compiled, AFRD coordinated with the Atlanta police and building departments and developed a plan of action for the glass hazard. The plan involved the following components:
- Atlanta Police ensured the closure of hazardous areas to pedestrian and vehicle traffic using hard barriers and officers posted in key locations.
- The fire inspection chief, a building department official, and the police department maintained contact with building managers to track the progress of removal of the damaged glass.
- The mayor’s office of communication shared the list of buildings with glass hazards and the affected streets with the news media.
Logistics. The department had recently developed the SOC Tactical Support Unit, which carries extra equipment such as saw blades and respirators. The ability to move support equipment quickly proved extremely beneficial during this incident. We must increase our capability to provide scene lighting. Light towers are valuable, but this scene required more portable lights than we could effectively provide.
Personnel. The Atlanta Fire Rescue Department does not operate a dedicated rehab unit. Each company carries a cooler of water with cups. This system proved woefully inadequate and was overwhelmed quickly. We must consider ways to increase our ability to rehab our personnel.
Many firefighters carry rechargeable personal lights, which offer several advantages for day-to-day operations. Several firefighters experienced lights with drained batteries during these operations. Most were prepared to continue operating with a backup flashlight. We should encourage personnel to foster the Boy Scout attitude, “Be Prepared.”
An operation of this magnitude will strain the resources of any agency. Feasible plans must be in place to create a unified command system and request mutual/automatic-aid resources. While handling the issues of the storm, the fire department still had to handle the standard day-to-day operations call volume. This type of operation can be expected to last up to five days. Even after the injuries are dealt with and the initial storm-related response has ended, countless issues arise regarding public safety and infrastructure as the city gets back to “normal business.” Maintaining a unified command post will help to deal with these issues in a more efficient manner.
It is also important to note that because of the extent of the damage to key buildings that are a part of the economic engine for the city, repair work will be a 24-hour-a-day venture. Much of the work is high-hazard because of the type of damage. Conduct site plans of the work zones, and develop contingency plans for response. As an example, the week after the storm, rappel and scaffold work in the city increased tenfold. The many broken windows in numerous structures pose a large fire hazard as well. Closed streets will increase response times dramatically and delay the arrival of the first company. Departments must be wary of dropping their guard too early.
The lessons learned during this incident were so many that we could not write all of them in the context of one article. It is easy for departments to become overly dependent on special operations personnel. During this storm response, engine and truck crews searched the vast majority of structures. Personnel trained to conduct technical rescues are needed in today’s modern world, but basic engine and truck work will save more lives than all the rescue teams put together. Firefighters must share these lessons learned with newer members of the department so that the learning continues to move forward.
CHRIS WESSELS is an assistant chief with the Atlanta (GA) Fire Rescue Department and a shift commander on the B platoon. A 21-year veteran of the fire service, he has served as an FDIC H.O.T. instructor since 1997.
STEVEN P. WOODWORTH is a battalion chief with the Atlanta (GA) Fire Rescue Department, assigned to Special Operations Command. He is a captain with the Fayetteville (GA) Fire Department and an instructor with Safe-Ir Inc. A 23-year veteran of the fire service, he has served as an FDIC H.O.T. instructor since 1997.
