Digital Extra: The Station Fire

A Providence Journal computer analysis of the fire that destroyed The Station nightclub last February suggests that even minor modifications to the building would have allowed dozens of more people to escape.

In one instance, the computer simulation found that relatively simple structural changes might have gotten everyone out alive.

The analysis, which used building-evacuation software called Simulex, incorporated the circumstances and structure of the West Warwick nightclub as they were on the night of the Feb. 20 fire, based on the accounts of more than 200 survivors, public records and a video made by a television cameraman.

The analysis was based on a crowd of 432 -- the number of people The Journal has confirmed were inside the club when the fire started. Although the software cannot predict the precise movements of specific individuals, it does illustrate the general patterns of crowd movement.

The software provides a tool to predict what would have happened had conditions been different, and gives a unique perspective on the events inside the club in the first minutes of the fire.

The analysis revealed how three factors combined to make the Station fire the deadliest in Rhode Island history: the time needed to escape the rapidly moving fire fueled by polyurethane foam, the placement and configuration of the exits, and the size and the density of the crowd.

The Journal analyzed how differences in the building exits and the size of the crowd might have changed the outcome.

Time Time was precious for those fleeing The Station, even more so because the fire, fueled by highly flammable polyurethane foam, spread unusually fast.

Interviews with survivors, the computer analysis and a video of the fire suggest most people who escaped had made it out in about 1 minute, 30 seconds.

Brian Butler, a Channel 12 cameraman who was shooting video for a story on building safety, walked out the main entrance about 1 minute, 7 seconds after the fire started. His video shows that light smoke had already reached the front door.

About 1 minute, 12 seconds after the fire started, the audio of Butler's tape picks up screams of agony from inside the club.

At about 1 minute, 25 seconds, the video shows two people emerging from a greenhouse-style window near the front door. Heavy black smoke pours from the window.

At about 1 minute, 36 seconds, Butler's camera captures fallen people wedged in a pile at the front door.

By about 2 minutes, 6 seconds after the fire started, the thick black smoke is streaming from nearly every opening in the building. People are still fleeing through doors and broken windows.

The Journal's analysis suggests that an orderly evacuation of the 432 people -- without the effects of heat, smoke and panic -- would have taken 2 minutes, 22.7 seconds.

That is 53 seconds more than the fire gave them.

That analysis uses only the doors, not the club's windows, as exits.

Exits The analysis suggests several dozen people were trapped well inside the club, before they were able to leave the main concert room and squeeze through the ticket-booth area and the first of two sets of doors on the route to safety.

"It slows down because you have to make a turn," Paul Wertheimer, a crowd-safety expert based in Chicago, said last week. "The ideal exit way leads you straight out."

Many survivors interviewed by The Journal said the area was so congested that they sought other ways out.

Attorney General Patrick Lynch has refused to release documents showing where bodies were recovered at The Station. The Journal has filed suit for access to the information.

The bottleneck suggested by the analysis trapped dozens of people near the center of the building, close to the heat, flames and poisonous smoke of the rapidly spreading fire.

This jam was before people piled up at the front doors, an event that was captured on Butler's videotape -- and broadcast around the world.

The analysis suggests that removing the ticket-booth area and the walls of the entry hallway at the main entrance would have taken about 35 seconds off the time needed to escape -- allowing everyone to get out in 1 minute, 47.4 seconds.

Patrons would then have had a direct route to the front door, without having to make a turn into the hallway.

In another scenario, the Journal analysis suggests that a double-door exit at the rear of the nightclub, by the main concert area, would have reduced the time needed to escape by about 38 seconds. That would have given the 432 people 1 minute, 44.7 seconds to get out.

"You want to split up the crowd," said Wertheimer. "It allows them to disperse quickly. It's the difference between a funnel and a spaghetti drainer. You don't want them all going to one side."

A more complicated modification in the analysis -- removing the ticket booth and entry hall, adding a bank of four doors at the rear, widening the front entrance to three doors and eliminating a low half-wall that separated two areas of the concert hall -- would have taken more than 60 seconds off the time needed to escape.

That would have allowed evacuation in 1 minute, 14.8 seconds, a time that would have given everyone inside the club a chance to escape.

Crowd Size While the Journal analysis suggests that 432 people might have gotten out in 2 minutes, 22.7 seconds, under ideal conditions, the newspaper also examined the time needed for smaller crowds to escape:

With 404 people -- the highest occupancy limit set by West Warwick in 2000, which required removing furniture from the club and hiring a uniformed firefighter -- the evacuation would have taken 2 minutes, 19.6 seconds, about 14 seconds after thick black smoke is pouring out of most of the building, ccording to Butler's video. There was no firefighter on duty at The Station.

With 317 people -- a limit set by the town in 1999, which required removing some furniture -- the evacuation would have taken 1 minute, 56.4 seconds.

With 258 people -- the limit if the town's conditions attached to having 404 were not met -- the evacuation would have taken 1 minute, 31.4 seconds, a time that would have given all but a few a chance to escape.

"There's still too many people in this venue -- legally or not," Wertheimer said, based on the Channel 12 video, which showed people shoulder-to-shoulder and more than a dozen rows deep in front of the stage.

Because people in the middle of a dense crowd cannot move until those toward the outside do, it takes a densely grouped crowd much longer to get going than the same number of people spread more thinly, according to Wertheimer.

This can be seen in the analysis done by The Journal.

In the scenario for 258 people, 75 people escape within 38 seconds of the fire beginning.

Scenarios with more than 258 people show fewer than 75 people escaping within 38 seconds.

The code Under the state's fire code, the four exits at The Station would have been enough, according to William Howe, chief of inspections for the state fire marshal. Howe generally discussed the code without specifically commenting on The Station.

One exit, a single door next to the stage, had a sign on it barring the general public, according to survivors. Through interviews, The Journal identified 20 survivors who came out that door.

Restricting exits is allowed under the fire code, according to Howe, as long as those exits are not counted when calculating the capacity of a building.

A second exit, a single door in the kitchen, was hidden and unknown to the general public. The Journal identified 12 people -- including four Station employees -- who got out there.

A hidden exit is allowed under the code, according to Howe, but, again, such exits cannot count in calculating capacity.

The third exit, a single door in the main barroom, was available to anyone who had noticed it.

About 46 survivors escaped through the barroom door, with an additional 54 going out the windows from the barroom.

The fourth exit, double doors at the main entrance, required fleeing customers to pass through the enclosed ticket area, an inner set of doors, the entry hallway, and then the outer set of doors.

The Journal identified 90 survivors who got out through the front door.

The Journal also counted 25 survivors who escaped through greenhouse-style windows next to the front door. There were no exit doors in that part of the club.

Under the fire code, standard-width doors such as those at The Station would be expected to allow 160 to escape, according to Howe. He said the capacity formulas are based on experience, and are not tied to evacuation times.

Ignoring the stage and kitchen doors, The Station had three doors available, one in the barroom and the double doors at the front. That meant, according to the formula in the state fire code, 480 people should have been able to get out safely.

The formulas do not take into account that a nightclub's walls might be covered with flammable polyurethane foam, which the fire code bans.

Said Howe: "The code probably assumes everything else is done correct."

In July, as a reaction to The Station tragedy, the state enacted broad changes in the fire code, which will take effect on the first anniversary of the fire. Under those changes, The Station's exits still would have been deemed adequate because the formula for calculating exit capacity does not change.

The new law will require automatic sprinklers in buildings like The Station. Sprinklers would be expected to give people more time to escape, but it is difficult to predict how much.

The software Simulex, the software used for The Journal's analysis of the evacuation, was developed by Integrated Environmental Solutions Ltd., a company in Glasgow, Scotland, that specializes in computer technology to help design and operate buildings. The software was developed in collaboration with researchers at Edinburgh University, in Scotland, and Lund University, in Sweden.

Simulex tracks each person's position in a building and calculates how fast that person walks, based on normal walking speeds and how close other people are.

Simulex accounts for differences in body size and walking speeds between men and women, as well as the effects of drinking alcohol.

Data for the analysis were gathered as part of an effort by The Journal to reach as many fire survivors as possible.

On the night of the fire and in the days and weeks following, Journal reporters interviewed survivors at the scene and elsewhere throughout Rhode Island, Massachusetts and Connecticut.

Eventually, more than 60 reporters were assigned groups of names to track down, based on a list that included names of survivors that had appeared in media reports and other documents, including a roster of survivors distributed at the state's family-assistance center in the days after the fire.

In September, The Journal published a list of 412 people who were in the fire. After the list was published, readers contacted the newspaper with tips that led to 20 more survivors being added to the list.

In all, The Journal interviewed 208 survivors. Among the questions they were asked were where they were inside the building when the fire broke out, how much they had had to drink, who they were with and how they got out.

The answers to these questions, and others, created a profile that was fed into Simulex before running the various evacuation scenarios.