On The Job - Alabama: Trench Collapse

On Sunday, May 24, 1998, the Dothan Fire Department was called to a tragic construction accident. A worker installing a sewer line was trapped beneath nine feet of earth when a "shear wall" collapse occurred. He and another man were working on the line over the Memorial Day holiday weekend to complete a project that would tie a new shopping center into the city's sewer system.

About The Trench

The trench was about 120 feet in length, two feet wide and nine feet deep. The earth had been opened for a few days before the cave-in and the area was in mild drought conditions. The last reportable rainfall had occurred 20 days prior to the incident. The very dry soil added to the danger of the operation and the instability of the trench.

It is difficult for the average person to comprehend, but soil is a heavy material - in fact, it weighs more than 100 pounds per cubic foot. Consider that a single cubic yard of soil (measuring three feet by three feet by three feet) contains 27 cubic feet of material, weighing in at over 2,700 pounds. A rule-of-thumb measurement is that a cubic yard of space is about the size of a typical full-size automobile trunk. After reviewing these facts, it is easy to see that the material at a trench cave-in has a tremendous amount of weight and, therefore, force.

Soils are classified by types, ranging from A (most stable) to C (least stable). The classification system is used to help determine the likelihood of a collapse and degree of difficulty of working with a particular soil. A measurement of unconfined compressive strength is used in determining these classifications. In layman's terms, under how much pressure will the soil crumble and break apart?

• Type A soils are cohesive with an unconfined compressive strength of 11/2 tons per square foot or greater. Examples are clay, silty clay and clay loam.
• Type B soils are cohesive and have an unconfined compressive strength greater than a half ton but less than 11/2 tons per square foot. Examples are angular gravel, silt or silt loam.
• Type C soils are cohesive soils with an unconfined compressive strength of a half ton or less per square foot. Examples are sand, gravel and sandy loam. Soils that have been previously disturbed, or which water is freely seeping, or that have collapsed are classified as Type C.

Trenches can collapse in three ways: lip slide, wall slough ("belly-in") and side wall shear. The lip slide is usually caused by piling the excavated spoil too close to the edge of the opening. This is the most common type of collapse and can be shored relatively easily. The wall slough is caused when a large section of soil falls from within the side of the trench. This may leave an overhang of material, which is extremely dangerous. In this case, the remaining uneven surfaces are difficult to shore. The side wall shear occurs when an entire section of the earth shears away and collapses into the trench. This is a dangerous situation because of the amount of material that has been disturbed.

The edges around the opening are called the trench lips. The removed material from trench is "spoil." The spoil pile should always be placed at least two feet from the lip because of the pressure that is exerted on the wall of the trench. Also, the spoil pile can fall into the opening and injure workers below. The "belly" is the side walls and the bottom is called the trench floor or toe.

During a cave-in operation, proper terms should be used to ensure understanding when communicating with construction workers. In fact, one of the first steps to resolve a trench incident is a simple drawing of the opening in question. While the drawing is being sketched, add the proper terms to avoid confusion. The sides of the opening should be uniquely identified, such as side A.

The Operation

The initial action plan was simple, but effective: evacuate the caved-in area, deny access to the hot and warm zones, size-up the situation, develop a victim-survivability profile, and request additional personnel and specialized equipment.

Police evacuated the area, curtailed all vehicle movement within 100 feet of the cave-in site, and provided perimeter security. Their biggest challenge was to stop the backhoe operator and bystanders attempting to rescue the trapped man. The equipment operator wanted to use the bucket of the device to unearth his partner, but that option was quickly ruled out by command.

The first concern was that the weight, vibration and movement of this piece of heavy equipment would cause additional collapse. Further, if the backhoe was able to be used without the fear of a secondary collapse, the next concern was to prevent dismemberment of the trapped worker, who was last seen standing at the bottom of the trench. (As a firm rule, all heavy equipment must be shut down and not used during a cave-in incident. The closest that heavy equipment should be operated is 100 feet from the collapse. This requirement holds true for earth-moving equipment as well as fire apparatus.)

As luck would have it, there was a lumberyard just across the street from this incident. Within a few minutes, we were able to have a large volume of lumber on location. Sheets of plywood were placed around the entire trench lip area. These sheets served as ground pads. They distributed the weight of the responders much like the action of ground jacks on an aerial ladder.

Next, the heavy equipment operator would be needed at the command post to provide information. About seven bystanders were digging with a few shovels upon our arrival. We asked them to stop because of the danger involved and to move to a safe area. The bystanders, however, made many comments that they were going to save a life and wanted the fire department to get out of their way. Every attempt was made to clear the area professionally, but because of the high emotions of the bystanders this was difficult. With these immediate needs handled, the incident commander was able to focus on size-up and the victim profile. The backhoe operator was now at the command post; he was our best source of information.

A simple drawing was developed that helped to determine the location of the trapped worker. About 22 feet of the 120-foot trench wall sheared on side "A" or the west wall. If the backhoe operator's information was correct, the victim was close to the middle of the cave-in, at the bottom of the trench. The heavy and tactical rescue unit was on location at about 11 minutes into the incident. Shoring panels were unloaded and a second load of lumber requested. Soon shoring panels would be placed and the most difficult job of hand digging would begin.

The mercury spiked at 97 degrees Fahrenheit with a relative humidity reaching 67%. The heat-stress index indicated that the temperature would feel like 124F. Further, the same chart cautions that an additional 10F must be added for any work performed in direct sunlight. Because there was no shade anywhere near the cave-in site, the temperature equaled 134F from a personal stress standpoint. Concern for firefighter rehab, rehydration and work/rest cycles were not far from the incident commander's thoughts throughout the incident.

The initial work/rest cycle was established at 15 minutes in the hot zone, either digging or building supplemental shoring. After a few cycles, the workload was simply too much to handle. Members were drained and on the verge of collapse due to the effects of the heat. The cycle was adjusted to 10 minutes of all-out digging followed by 20 minutes of rest and rehydration. This balance seemed to work well and was enforced until the sun set. The work time was then extended to 15 minutes due to the lower temperature.

Because of the vast amount of earth that had to be moved by hand, a large number of responders would be needed. As well, an elaborate support system had to be developed to keep them working. Our personnel accountability system had a tremendous workout, but was able to track all 145 people throughout the incident. The outside agencies were the accountability challenge - they were added to the passport system using blank name tags and a grease pencil.

Further questioning of the backhoe operator provided more details about the opening. There were no pipes, utilities or other obstructions in the ditch. This was bad news, because it dashed hopes that the trapped man might be protected in a void space. However, it did rule out potential problems with leaking gas or water or electrical hazards. Lock-out/tag-out was not required in this case. As a precaution, however, the water line nearest the trench was shut off.

The six sheeting panels (three per side) were placed and installed into the opening. Each panel had to be secured by trench jacks and piping that was cut to size on location. The trench jacks were then cleated to make sure that they would not fail during the hand-digging operation. About 12 members who are rescue technicians were used in the more difficult tasks, while the members who do not have this specialty training were used in more general capacities.

Two members were assigned to dig. The "entry team" wore rescue harnesses with a lifeline attached. The lifeline had one member assigned to belay the rope outside the hot zone. For each person in the trench digging, we had one person that would pull up the full plastic bucket and send down an empty one. Once the dirt was at the topside of the opening, another person would carry away the spoil to a remote pile out of the way.

Positive-pressure ventilation (PPV) was activated by using an electrically powered smoke ejector. A directional shroud was attached to the discharge side of the fan and directed into the hole. This action was an attempt to provide improved air quality for the victim and help the members assigned to dig. Two ladders were used with lumber added to allow for safe crossing of the trench. Two more ladders were placed into the now shored opening at the ends for access, egress and escape routes.

The ongoing action plan was continually updated. The plan included keeping security over the hot and warm zones, providing support for the additional personnel, establishing a formal rehab sector, handling logistical needs and updating the media. Many people gathered at the incident and wanted to gain access to the caved-in section. Permission was denied and police maintained security.

The rehab sector split the roof cover that protected the staging area. Fire-line tape was used to separate these functions. The rotations for the crews were from the operations section to the rehab sector, then to staging and back to operations. Each time a person was moved in or out of the operations section, the passport system was updated. Police were assigned the public information function and gave media updates.

Because of the large number of people needed to complete this operation, outside mutual aid was requested early into this alarm. The Fort Rucker Fire Department sent 10 on-duty members; 16 volunteer members from Houston County fire stations answered the call.

Upon arrival, units were sent to a staging area that was under roof, which provided much needed shade. This helped to prevent or minimize the heat stress of the troops before they could be deployed in the operations section. By now, 30 off-duty members had responded to a callback request. As well, all 45 on-duty members would be used at various times at this operation. Members were moved in teams of four from staging to a waiting area just outside of the hot zone. As the shoring or trenching sector needed people, they were activated within seconds. The operations area was covered as well as the waiting area with pop-up tents. Again, this action was an attempt to prevent heat-related injuries to the emergency workers.

As the operation continued, the digging crews would reach the bottom of the shoring panels. At this point, supplemental shoring would have to be built on the spot to fit the exact opening. Supplemental panels had to be added three times before we could reach the trapped construction worker. As day began to turn into night, the air-and-light unit was readied. Lighting would have to be provided in the trench and around the work area.

After nearly 11 hours of digging and building shoring, we extricated the victim. Unfortunately, the man was completely engulfed in dirt and covered by tons of material, and did not survive the accident. His remains were released to the police and Houston County coroner. The autopsy indicated that the man was able to survive only a few minutes and died before the 911 call could be placed.

Efforts now focused on supporting the accident investigation by police and the Occupational Safety & Health Administration (OSHA). Information was gathered throughout the incident and turned over to the OHSA inspector.

The last phase of this operation was to develop and implement a demobilization plan. Every company in the city had rotated through and equipment recovery would take several days. The trench had to be secured to prevent others from falling into the opening. Lighted barricades were brought in from the public works department to cordon off the area safely. Equipment and materials had to be returned to various groups.

Lessons Learned Or Reinforced

1. Family members of the victim arrived on the scene about two hours into the call. Chairs were set up in the cold zone for them and they were given updates by ranking officers. A priest also was called in. When the trapped man's wife was initially briefed, she fainted and was taken to a local hospital. Having the family watch our operation put extra pressure on us. In retrospect, we should have provided a better facility for the grieving family, perhaps in a small city bus or van. That action would have taken them out of the heat and noise and away from the situational stressors at the scene.
2. It would have been more efficient to have separated the rehab and staging areas. Having them together made it difficult to keep track of the rest time of the firefighters. Crews would report ready status before they had proper rest and re-hydration.
3. The department is considering incorporating a vehicle that can store various sizes of lumber and related materials. A major part of our success relied on the fact that the lumberyard was open and nearby.
4. The fire department needed to assign a public information officer (PIO) to work with the police. The media wanted detailed and specific rescue information that was not available from the police PIO.
5. The fire department needs to develop a prevention program that addresses the issue of trench safety. At most anytime in our city, improper trenching operations are underway. To prevent these tragedies would be the best response to any emergency.

The Epilogue

As mentioned earlier, a small but vocal group had started a rescue attempt prior to the arrival of the fire department. When these people were asked to leave the caved-in trench, tempers flared. A fire captain was told that he did not have the authority or the "strength" to remove these would-be helpers.

Little did I understand the grief that would soon be facing me and the department. The group brought a family member to the warm zone (slipping through the fire-line tape) to show just how slow we were working. The comment was made that they could have had the victim out, if only allowed by the fire department to do so. After working to explain to the family members how the operation would be handled and build a level of confidence, this trust was quickly broken. During the rest of the incident, the family members were skeptical of any information that we shared with them.

The next attack the disgruntled "rescuers" made was with a local TV station. To my disbelief, a news story that aired the next day contended that the fire department refused the help of this band of rescuers. The story went on to say that the time to extricate the victim was too long and that the fire chief refused to use a nearby backhoe to save the man's life. The clip ended by saying that the governing body of the city would be holding closed-door meetings to discuss the rescue operation. In fact, the city commission had no such plans and the city manager was very complimentary about our operation. The story was aired twice and caused a strong reaction from the firefighters that had just given their all at a very difficult alarm.

A reporter from the local newspaper showed up at the incident critique. Reporters are not invited, as a rule, so that open discussions among responders may be facilitated. In this case, I let the reporter stay and participate.

Believing that the attacks were unsuccessful and had subsided was wishful thinking. I then was contacted by one of the family members. The dead man's brother had been called by the rescuers group and told they "could have helped and saved the trapped man's life." They encouraged the family to file suit against the city, the department and the fire chief.

The reason why I am sharing this unpleasant information is because it is easier to learn this lesson from somebody else's experiences. It would be easy to take the path of least resistance and give in to the naysayers. If I were to do that and not defend or justify our operation, the next time we would cut a corner or two this practice would be paid for by the backs of the "rank-and-file" members. This is one battle worth fighting and I would urge all fire chiefs to do the same.

The learning curve has been steep. The first lesson is, even in the heat of battle, that the untrained people who get in our way are still customers. The removal process should be as professional as possible. In this case, after about 15 requests, the would-be rescuers listened reluctantly. Next, there is a clear need in our community to develop a 15-minute presentation that discusses trench rescue. This program will highlight how to prevent cave-ins.

The last lesson deals with contact with the media. We should have stated our side of the story, rather than letting the police, an agency without specific rescue expertise, do so.

Dothan Fire Department
May 24, 1998 Trench Rescue

1145 hours - Initial alarm. Paramedic Engine 9.
1147 - Special alarm. Truck 25.
1148 - Fire chief and Battalion 1 self-dispatched.
1148 - Special alarm. Squad 1.
1151 - Special alarm. Paramedic Engines 4 and 5.
1153 - Special alarm. Trench Rescue Team.
1157 - Special alarm. Paramedic Engine 1.
1159 - Special alarm. Truck 5.
1214 - Special alarm. Rehab 1.
1219 - Special alarm. Private ambulance company 1.
1229 - Emergency medical services self-dispatched.
1301 - Fire Prevention 1 self-dispatched.
1309 - Special alarm. Engines 23 and 24.
1313 - Fire department chaplain self-dispatched.
1324 - Special alarm. Paramedic Engine 8.
1338 - Fire Prevention 2 self-dispatched.
1340 - Special alarm. Houston County volunteers.
1352 - Special alarm. Fort Rucker Fire Department.
1447 - Special alarm. Light-and-Air Truck.
1634 - Special alarm. Houston County Coroner's Office.
1842 - Special alarm. Private ambulance company 2.
1849 - Special alarm. Private ambulance company 3.
1933 - Special alarm. Engine 6.
1958 - Special alarm. Engine 25.

RESPONDING AGENCIES

Dothan Fire Department (68 personnel), Dothan Police Department (19), Dothan Fire Department chaplains (2), Dothan Fire Department Ladies Association (2), Houston County volunteers (15), Pilcher's Ambulance Service (6), Houston County Emergency Management (2), Fort Rucker Fire Department (10), Salvation Army (8), MAST helicopter crew (4), Dothan Water Department (4), Dothan Public Works Department (1), Southeast Alabama Regional EMS Office (1) and Houston County Coroner's Office (2).

Author's note: Special thanks to Captain Keith Gaster for help with the diagrams and to Intern Matt Martens (OSU) for his research.