D.C. Fire Chief Kenneth B. Ellerbe is seen at the Wilson Building in Washington on March 28. Data obtained by the Post raise questions about the fire department’s arson clearance rate. (Marvin Joseph/The Washington Post)

The D.C. fire department has quietly changed the way it counts arson fires, resulting in a dramatically lower arson rate for the city and a dramatically higher case-closure rate for the department’s investigators.

Interviews with fire investigators, department officials and national experts, along with a review of internal documents, cast doubt on data the fire department has reported on the number of arsons in the District and its success in closing those cases with arrests. They also reveal a sharp disagreement between Fire Chief Kenneth B. Ellerbe and his fire marshal over the method used to count arsons.

At a D.C. Council hearing April 17, Ellerbe was asked how his investigators had been able to solve arson cases 72.7 percent of the time — more than three times the national average.

“What accounts for the improvement?” council member Tommy Wells (D-Ward 6) asked. “Has there been any change of what you’ve been measuring?”

Ellerbe said there had “not been much of a change” in measurement.

“I think that as employees gain more and more experience, they become more and more successful at what they do,” Ellerbe said. “In addition to that, I believe the rate of arson cases went down.”

But the department had changed the way it counted arsons.

At the hearing, Fire Marshal Bruce D. Faust alluded to the change, telling Wells that he had compiled a “very different” arson closure rate. Under Faust’s calculations, contained in a recent report to the D.C. Office of the Chief Financial Officer, the department recorded a 9.6 percent closure rate for fiscal 2012.

Previously, in keeping with general practice across the nation, a fire was considered an arson when investigators determined that it had been deliberately set. But sometime in recent years, the department narrowed its definition of arson cases to those in which there was evidence of willful or malicious intent sufficient to support an arrest, according to interviews with department officials.

The change had a dramatic effect: The number of reported arsons dropped by nearly 80 percent, from 154 in fiscal 2009 to 32 in fiscal 2012, according to data from the Office of the City Administrator. As that number fell, the department reported that its success in closing arson cases had surged to 72.7 percent.

Department officials later explained that the 72.7 percent figure was based on partial data: eight arrests out of 11 arsons through May 2012. The full-year rate was 34 percent, they said. Still, the higher figure was presented to the council in February and included in the department’s 2014 budget proposal.

Fire marshal’s dissent

The District’s arson numbers and arson closure rate were called into question as early as last summer by Faust, a deputy chief who works for Ellerbe and oversees inspections and investigations.

Faust calculated the 9.6 percent closure rate for 2012 using an older, more universally accepted method: He divided the number of arson arrests, 11, by the number of “incendiary” fires, 114. (He did not count incendiary vehicle fires, because under D.C. law those fires are not subject to arson charges.)

The statistical duel comes at a difficult time for Ellerbe, who has been embroiled in controversy over emergency response times, rebuked by a union’s no-confidence vote and questioned about the reliability of information the department has provided during his watch. Last month, Ellerbe apologized to the council for submitting inaccurate data on the fire department’s fleet of vehicles.

Wells, chairman of the council’s Public Safety Committee, warned Ellerbe at a March hearing that he should be prepared to “put [his] job behind” the accuracy of the information the department provides.

The department made no public announcement about a change in the arson standard, and Ellerbe did not refer to it when questioned by Wells. Fire officials described the change after receiving inquiries from The Washington Post about the different closure rates, but they did not answer questions about who mandated the change or the precise nature of the new classification system.

Faust has raised concerns in the past about the department’s failure to have a consistent standard for counting arsons.

“Over the years, there has been much debate over arson closure rates and how to determine the rate, what to include and what not to include when calculating the number,” Faust wrote in a memo last summer to the D.C. Office of the Chief Financial Officer, which was obtained by The Post through a public-records request.

Inconsistencies, he wrote, could render the rate “meaningless” or “misleading.”

“I don’t think there’s any validity to it,” Faust said of the reported arson closure rate in a recent interview with The Post. “I don’t think there’s any value.”

The rate, Faust said, incorrectly suggests that arson investigations have drastically improved in recent years, even though the number of investigators has fallen by nearly half during Ellerbe’s tenure and the number of arson arrests is at an eight-year low.

The standard definition

In the past, the department calculated the closure rate using a definition of arson as an “incendiary fire,” or a fire “deliberately ignited when the person knows a fire should not be ignited.” That definition is found in the National Fire Protection Association’s handbook for investigations. The number of closed cases was divided by the number of incendiary fires to determine the rate.

Former D.C. fire chief Dennis L. Rubin, who served until the end of 2010, said incendiary fires were generally reported as arsons during his tenure. He said he would not question the analysis produced by Faust, who was also Rubin’s fire marshal and “knew his job inside and out.”

Asked why the department would report numbers different from those used by its fire marshal, Ellerbe’s chief of staff, Andrew Beaton, said Faust should report arson data the way he sees fit.

“I think the data that’s submitted to the city council should be very accurate to what the fire marshal wants to do in terms of the calculation,” Beaton said. He added: “Look, the information is what it is. I’m not trying to reformat it in any way that makes it look better or worse.”

Sometime after Ellerbe took over in 2011, the department stopped reporting all incendiary structure fires as arsons, an analysis of internal data shows. The result has been an arson total that appears significantly lower than the national average for cities of comparable size.

Cities with populations of 250,000 or more report an average of 96 intentional structure fires a year, according to a 2009 arson analysis by the National Fire Protection Association. That is three times the 32 reported by the District last year.

‘Impossibly low’

“I can give you neighborhoods in Miami that have 32 arsons,” said Sandy Burnette, a nationally known arson expert and lawyer who practices in Florida. “That level of arson in the District strikes me as inordinately low and, quite frankly, impossibly low.”

Beaton said arson numbers in recent years may seem low because the department is counting only “proven cases of arson,” cases that have found their way into the legal system. “They’re not the same” as suspected cases, he said.

Fire department spokesman Tim Wilson said that in recent years, the department has drawn a distinction between “arson” and “incendiary” to “reflect a more detailed analysis of how arson cases are identified and closed.” The department now considers “arson” to define a “willful, malicious intent to start a fire” and “incendiary” to define “an action or method used to start a fire,” he said in an e-mail.

The new definition of “incendiary” does not conform to language in the National Fire Protection Association’s guidebook for investigators. Burnette, the arson expert who practices in Florida, said he was puzzled by the D.C. fire department’s explanation.

“I know of no such distinction between incendiary fires and arson fires,” he said. “I have always considered those to be one and the same.”

Wilson, the department spokesman, did not say how the department came up with its new definition of incendiary fires. He did say that the department considers the District’s arson law, which refers to “malicious burning,” to be a more precise guideline for how to count arsons.

A ‘gray zone’

John Hall, who works for the National Fire Protection Association, said a “gray zone” exists in interpreting fire investigation guidelines. “Unfortunately, the landscape on several critical points here is not that bright and sharp,” Hall said. “The city deserves the benefit of the doubt. If they cross the line and try to compare apples and oranges, or compare numbers by another definition, then you can call foul.”

In an e-mail, Wilson said the change in the department’s way of counting arsons also explains a discrepancy in data reported for fiscal 2011. The department reported 99 arsons for 2011 in an annual report delivered last year to the council. But in an oversight report delivered in February, the number was cut to 23, a 77 percent decrease.

Wilson did not respond to a question about why Ellerbe did not tell the council committee on April 17 that his department had changed the way it counts arsons.

Ellerbe agreed to an interview last week to discuss arson investigations.

When asked the first question, about how the fire department calculated its arson closure rate, Ellerbe motioned for one of his captains to respond, saying, “These are questions for our fire investigators, right?”

Capt. Tony Falwell began to flip through pages of an internal spreadsheet, showing the chief how the data was organized.

“Relax,” Ellerbe told Falwell. After several seconds of silence, Ellerbe stood up and walked out of the room. He did not return.