100 souls. That’s the number of sons, daughters, moms and dads lost in just 90 seconds during The Station nightclub fire on Feb. 20, 2003. In addition, more than 200 people were injured.

Despite several famous theater and nightclub fires throughout our nation’s history before 2003, many innocent people died senselessly. 

Why? What happened to cause such a large loss of life so quickly?

And most importantly, what can be done to prevent another similar tragedy?

The Show

At 11 p.m., approximately 440 people packed into The Station, a wooden construction, one-story building known for its live music shows, to rock out with the well-known band Great White. As the band started its set, pyrotechnics were ignited near the stage, sending a shower of sparks upward and toward the walls. 

Highly combustible polyurethane foam affixed to the walls caught fire almost immediately and acted as kindling to the adjacent wooden wall surfaces (paneling). Although evacuation began within 24 seconds of the fire starting, a combination of factors resulted in crowding at the doors, trapping many inside. 

Toxic black smoke and heat filled the club, reduced visibility and incited panic. The exits were quickly overwhelmed by the occupants rushing toward them. Within 90 seconds, the smoke, heat and human crush had claimed 100 souls, including Great White’s guitar player.

The Building

The Station was originally constructed in 1946 in West Warwick, R.I., as a nightclub. When built, building and fire codes were regionally developed and adopted, very different from the widespread use of model codes today. It is unknown if the original building or its many alterations were constructed in accordance with the myriad of codes and standards that have existed since the 1940s. 

As stated in the National Institute of Standards and Technology’s (NIST) “Report of the Technical Investigation of The Station Nightclub Fire,” at least four different construction codes may have been adopted. NIST’s technical investigation did not focus on compliance with the various adopted codes and standards.

Compared to model codes enforced in 2005, when the NIST report was written, the club had a number of code violations. Besides the use of pyrotechnics without a permit or precautions, both the combustible wall finishes (wood paneling) and the design of the building’s entrance were major contributors to the loss of life. 

The main entry had double doors that opened outward; however, those led to a small foyer with a single interior door. Although three other doors were available and marked with lit exit signs, nearly two-thirds of the occupants tried escaping through the main entry. The total exit capacity was inadequate for the number of occupants in the short time available. 

Finally, although the 2003 model building codes would have required fire sprinklers in the nightclub (Use Group A-2, occupancy over 300), they were not installed due to grandfather clauses for buildings constructed before 1976. A fire alarm system was installed and operated 41 seconds after ignition, well after occupants near the stage — the fire’s ignition point — were already moving toward the exits. 

Although the fire department was notified only 40 seconds after the fire started and was on the scene less than four minutes later, nothing the fire department could have done would have saved the building or the souls trapped inside.

No single issue led to the massive loss of life. Instead, the combination and accumulation of challenges over time resulted in the horrible outcome.

Reconstruction and Modeling

In an incredible coincidence, most of the events before and during the fire were captured on high-quality video by a local news camera. NIST used the footage to reconstruct the scene, egress and fire growth. NIST used both empirical evidence — including a partial rebuild and burn of the stage area, complete with pyrotechnics, foam and wood paneling — and calculations using NIST’s Fire Dynamics Simulator and Smokeview software to study the impact of various factors on the outcome of the fire.

The burn tests were used to “dial-in” the calculation parameters as well as gain insights regarding temperature, heat flux, oxygen and toxic gas levels. All these factors impact tenability, or survivability, during a fire.

Table 1 is a simplified version of the information presented in the NIST report. As shown in the table, all four of the tenability factors listed were indicated to be exceeded 87 seconds after ignition.

Of particular interest to NIST (and the author) was the relative value of alternative building designs and fire safety measures. In this case, a fire alarm system proved unsuccessful at saving lives. But could lives have been saved by increasing exit width to meet 2003 code requirements? Would fire sprinklers installed in accordance with typical NFPA 13 requirements effectively have controlled the fire growth?

What if More Exits Were Installed?

NIST used two egress modeling programs, buildingEXODUS and Simulex, to approximate occupant behavior and egress time in a number of scenarios, including “no fire” and a scenario where exits met the 2003 model building codes. 

The modeling software used the same occupant travel speed for all scenarios. The quickest egress time was the “no fire” egress scenario using the exact door configuration present in the building, requiring 188 seconds. 

Interestingly, NIST’s egress modeling showed that doubling the front door width to meet current codes resulted in increased egress time, up to approximately 196 seconds. The greater time for egress compared to the first scenario is because more digital occupants used the front door, the nearest exit, to escape even when the queue was longer than other options. 

This programming was selected based on actual behavior observed in the fire. Most occupants tried to leave out of the same door they entered rather than using other available exits.

At least 166 digital occupants remained inside the building in all modeled scenarios at the 90-second mark. Fortunately, during the actual fire, concertgoers escaped through broken windows, reducing the actual fatality count when compared to computed models.

If additional exits and earlier detection couldn’t save lives, would sprinklers have helped?

What if Sprinklers Were Present?

NIST’s physical reconstruction installed four quick-response, K5.6, ordinary temperature sprinklers in a 12-foot by 12-foot pattern at the stage and main floor area — plus a fifth sprinkler installed at the drummer alcove at the rear of the stage — as shown in Figure 2. Sprinkler spray density was approximately 0.10 gpm/ft² with all sprinklers operating.

During the reconstructed burn test, three sprinklers operated. Above the platform at stage-left, the first sprinkler activated in 23 seconds. The sprinkler above the platform on stage-right was the next to activate at 26 seconds. One second later, the sprinkler in the alcove activated.

The time to reach untenability is presented once again in Table 2 and now includes a comparison between the sprinklered and unsprinklered building reconstructions.

NIST’s reconstruction experiments of the fire at the stage demonstrated that “a fire sprinkler system installed in the test room in accordance with NFPA 13 was able to control the fire initiated in nonfire-retarded polyurethane foam panels and to maintain survivable conditions at head height in the test room for the duration (over five minutes) of the experiment. A computer simulation of the full nightclub with and without sprinklers led to a similar positive result for the sprinklered scenario.”

Lessons Learned

The NIST report recommended several proposals to improve life safety, such as enforcing existing code requirements. In addition, NIST’s report suggested many building code modifications that were ultimately considered by code committees. Not all recommendations were ultimately implemented, but there were two significant changes.

First, the 2006 edition of the International Building Code changed the threshold for mandatory sprinklers in Group A-2 occupancies from 300 occupants to 100 occupants. This threshold aligns with the occupant count for panic hardware in Group A occupancies. 

Second, although The Station was a nightclub, the reduced sprinkler threshold was extended to all Group A-2 occupancies, including restaurants that do not serve alcohol. As a result, even some large fast-food restaurants with large seating areas or play areas may be required to install sprinklers.

Clearly, had The Station been equipped with fire sprinklers, outcomes may have been much different for the 100 souls lost.

Christopher Culp, PE, CET, is the director of fire and life safety and a principal for Henderson Engineers. He helps manage Henderson’s 80-person fire and life safety discipline and has 25 years of experience at the firm in the design of fire protection systems and building code compliance. Culp has served as a principal for the fire and life safety group since 2007.