Firefighters were called to Metro stations at least 813 times for smoke- and fire-related incidents from January 2014 through December 2017 — calls that were most prevalent at stations along the western end of the Red Line, in the downtown core, and at the juncture where trains on the Orange, Blue and Silver lines enter the District from Northern Virginia, according to a Washington Post analysis.

Those numbers, compiled from data obtained from fire departments in the six jurisdictions served by Metro, are much larger than the transit agency’s official count of fires that occurred in the system. According to Metro’s tally, which is reported to the federal government, there were 312 fires during that period.

The disparity comes from the way Metro defines a fire: a blaze that “required suppression” or resulted in injury, evacuation or “substantial damage.”

The fire departments’ reports show that they respond to a variety of fire- and smoke-related disruptions on tracks and in stations, including sparking, smoldering and smoke incidents that don’t cause extensive damage, but require temporary single-tracking that disrupts travel and causes delays for riders.

The analysis also provides a glimpse into how the system earned a national reputation for being fire-prone.

The numbers from the fire departments, like those from Metro, do show that the frequency of fires in the system has decreased since the end of 2016 — probably because of factors including the year-long SafeTrack reconstruction project, reductions in service, a more intense nighttime track inspection schedule and a $12.2-million project to plug water leaks on the western end of the Red Line. Water leaking into the tunnels degrades electrical equipment and can cause fires.

The Post obtained fire department data from January 2014 to December 2017, though Montgomery and Prince George’s counties did not provide data from the last four months of that period, so the fire responses in those jurisdictions are probably undercounted.

Each department varies slightly in the criteria it uses to categorize events, but the numbers compiled here include incidents in which firefighters were called to Metro properties to respond to things such as a “fire alarm,” “smoke scare,” “rail vehicle fire” or “tunnel box alarm.” (A smoke scare, for example, could be a case of smoking brakes rather than a track problem, but the fire department is called regardless.)

The numbers do not include medical emergencies, reports of people being stuck in elevators or times when the call to the fire department was canceled while firefighters were en route. Redundancies in events have also been removed.

In response to inquiries about the data, Metro officials said the fire department logs do not accurately reflect the number of actual fires in the system. Additionally, they said, the higher number of calls is a sign that Metro is conducting business more safely and taking more precautionary measures when workers spot sparks or smoldering on tracks.

The number of times firefighters show up at a station, they said, is not a reflection of the state of the infrastructure.

“That’s not necessarily the best empirical metric of the rail system’s health,” Metro spokesman Dan Stessel said. “It is a demonstration to our commitment to safety.”

But the fire departments’ data provides not just the number of times authorities were called, but also insight into when and where those incidents occurred — details that are not publicly available from Metro.

And because a call to the fire department often means temporarily holding trains from an affected area, the data offers a sense of why so many riders feel burned by the system.

Katherine Kortum, chair of the Riders’ Advisory Council, said the numbers illustrate the depths of the system’s troubles at their height in 2015 and the lasting impact the safety crisis had on riders.

“The image lasts longer than the fires,” she said. “The fires are significantly reduced from what they were, but when customers or riders have experienced that, it’s hard to forget.”

Behind the numbers

More than 30 percent of the 813 incidents during the four-year period took place between Shady Grove and Dupont Circle stations.

That means that if you’re a commuter traveling between downtown and Shady Grove, you may have encountered as many as 253 smoke- or fire-related incidents outside of the downtown core. It also illustrates that when it comes to the disruptions caused by track fires, people commuting to or from the western end of the Red Line probably have it the worst.

Medical Center had the most, with at least 46 fire-related incidents logged from January 2014 through August 2017. After that was Bethesda, then Friendship Heights.

Beyond those, the cluster of stations with the most such incidents are downtown: Gallery Place, L’Enfant Plaza and Metro Center. Those critical junctures were affected by a report of smoke or fire 96 times. Incidents in these spots are particularly problematic because those stations are transfer points, and an interruption in service can have a ripple effect throughout the system.

In the list of the stations with the most smoke and fire incidents, Foggy Bottom holds the No. 7 spot, with Rosslyn at No. 12, highlighting the risk of meltdowns in and around the capacity-strained Rosslyn Tunnel, a chokepoint in the system.

The number of fires in the system escalated through the years — partly because of ailing infrastructure before the launch of SafeTrack in summer 2016 and partly because of increased reporting of potential fires by Metro employees after the January 2015 L’Enfant Plaza calamity.

On Jan. 12 that year, one rider died and scores of others were injured when an electrical arcing incident caused a tunnel outside L’Enfant Plaza to fill with smoke. A train stopped in the tunnel was engulfed in the toxic fumes.

After that, changes were made: A fire department official was permanently assigned to Metro’s operations control center, 24 hours a day, to help dispatch firefighters at the first sign of a spark. And Metro staff were instructed to take small events more seriously as a precautionary measure.

Subsequently, the running count of monthly fires jumped sharply: The total of fire-related calls to departments in December 2014 was 5. In January 2015, it was 19. And in February 2015, it was 32.

The following six months saw some of the highest number of calls to fire departments in the system’s history. In June 2015, there were 50 smoke- or fire-related incidents at stations throughout the system that required a fire department response.

One of the causes of these fires is water infiltration — rain or groundwater leaking through the concrete of the tunnel system, causing pools of standing water that can serve as an electrical conductor, or intensify the degradation of electrical equipment, causing metal equipment to smoke or spark.

In a recent analysis, Benjamin Ackerman, Zhicheng Ji and Jacob Fiksel, biostatistics PhD candidates at Johns Hopkins University, assessed whether there’s actually a relationship between rain and the number of times firefighters respond to smoke or fire calls at Metro stations.

Using historical records of precipitation, they determined that, while the few days with multiple smoke and fire incidents tended to have slightly higher amounts of rain, there were not enough of those days to draw definitive conclusions. Ultimately, they found no statistically significant relationship between precipitation and the odds of calls to firefighters for fire-related incidents on the Metro system.

How other systems compare

Metro’s fire statistics are not easily comparable with other subway systems.

While Metro is the country’s second-busiest, it is dwarfed in size by the New York subway, which is also more than a century older. The New York system, three times as big by track mileage, has exponentially more fires than Metro, according to National Transit Database stats tabulated in a 2016 Washington Post analysis, which found New York has 27 fires per million revenue miles to Metro’s 2.1.

Still, the data in that analysis predated Federal Transit Administration requirements that caused the number of fires Metro reported to more than double — from 49 in 2015, to 129 in 2016.

Still, the analysis showed that among its peers — the BART system in California and MARTA in Atlanta — Metro leads the pack.

Officials have blamed the New York system’s fire problem mostly on heaps of trash littering its tracks. Metro’s prohibition on food in the system results in substantially less waste, though debris still poses a problem.

The National Transit Database analysis showed that BART has 0.34 track fires per million revenue miles. By comparison, ­MARTA has 1.9 fires per million miles.

Fires on the century-old systems — New York, Chicago and Boston — far outpaced Metro in the analysis. Boston had 21 fires per million revenue miles, and Chicago had 7.4, according to data compiled by The Post. But with the new reporting requirements, Metro had as many or more fires than the Chicago system in 2016 and 2017.

In 2016, Metro had 129 fires to Chicago’s 79, and in 2017, Metro had 101 compared with Chicago’s 108, according to data provided by both agencies.

Impact on riders, firefighters

While problems with rail cars are the most common cause of subway delays, smoke and fires can be among the most costly in terms of riders’ time.

Fires take tracks out of service in both directions and often require a temporary shutdown of power to the third-rail in the affected area. They cause a chain reaction that snarls passenger trips far down a line. As trains stack up at rush-hour, platforms become like holding pens with limited capacity, giving rise to the conditions sometimes labeled “meltdowns.”

When stray current caused a shutdown of the Red Line between Gallery Place and Dupont Circle one morning in April 2017, riders reported waiting as long as 90 minutes aboard crammed trains.

The large number of incidents, the vast majority of which are relatively minor, also take a toll on the area’s fire departments. In 2014, Metro officially reported 33 fires — even though firefighters were called to Metro properties 95 times. In 2015, though Metro logged 49 fires, firefighters were called to stations at a far higher clip: 386 times.

After a change in reporting protocol, those numbers began to grow more aligned. Metro reported 129 fires in 2016, which was more than half of the number reported to the fire departments around the region — 223.

And in the first two-thirds of 2017, Metro logged 73 fires, while fire departments responded to 93 smoke- and fire-related calls at Metro stations.

D.C. Fire and EMS detailed the procedures that kick in every time they get a Metro call.

For example, every “box alarm” requires the same coordinated response: five fire engines, two ladder trucks, a rescue squad, three battalion chiefs and an ambulance unit. If the incident is in a tunnel between two stations, two additional engines and another truck are sent to a second station. Fire officials said the response ensures that firefighters are prepared for every possible scenario. 

“This is basically a high-rise building laid on it side,” D.C. Fire Chief Gregory Dean said in a 2017 interview. “You look at the amount of resources we send — same thing you do for a high-rise building. We send a number of resources until we can determine what’s happening.”

In the scope of all the types of emergencies that a department responds to in an average month, Metro’s recurring issues are not a significant drain on the department’s resources, he said. But they do require a very specific protocol.

Here’s what happens once firefighters arrive: Fire crews enter the station, prepared for the unlikely event that people need to be rescued or that a fire needs to be extinguished. Those crews have to scope out a number of factors: What conditions are being reported from the rail operations control center? What caused the alarm to go off? Do crews have to enter the track bed? Is the third rail energized?

Usually, they end up standing on the platform while a Metro worker checks the tracks, concludes that the sparking or smoke has dissipated on its own and gives the all-clear. All the while, passengers on nearby trains are waiting.

“Probably 9 out of 10 calls, we don’t go into the tunnel or even the track bed,” Assistant Fire Chief John Donnelly said in the same interview.

By September 2017, on the heels of the spike in calls from 2015, the fire department had rolled out a reduced assignment that required fewer resources: three engines, two trucks and two battalion chiefs, according to Donnelly. It’s a product, he said at the time, of increased reporting in an era when emergency calls come not just from Metro, but from Twitter-savvy riders and observers who have an awareness of potential risks. 

“With the advent of cellphones, with the advent of publicity and some of the new problems we faced, we had a lot of calls coming from outside of Metro that weren’t emergencies but were to investigate an odor of smoke,” he said.

By late 2017, the average time for D.C. Fire and EMS to respond and clear an incident was 38 minutes. From there, it’s Metro’s issue to handle.

Metro carries hundreds of thousands of commuters daily, and fires often break out at busy transfer stations such as Gallery Place and L’Enfant Plaza. So how many fires should a transit system be expected to have? 

“We think zero’s pretty good,” Dean said. “We put our people in harm’s way every time we respond, so we want to find ways to eliminate that as much as possible.”