Third in a four-part series
Pam McArthur-Elliott didn't like driving over the railroad tracks on Ferry Street in Anoka. The crossing gate sometimes malfunctioned, forcing her to find another way home.
On this clear Saturday evening in 2002, however, the gate was up and the lights were off. It looked safe. But just as McArthur-Elliott was crossing the tracks, she later testified, a Burlington Northern Santa Fe locomotive came roaring out of nowhere. She floored the accelerator in her Ford Escort.
"The gates never went down, they never went down!" said McArthur-Elliott, whose husband called 911 to report the incident right after she got home, still shaking. "Thank the Lord, I guess it wasn't my time."
Eighteen months after that close call, four young people weren't so lucky. They were killed when their car was crushed by a BNSF train at the same crossing, a heavily traveled intersection that generated more than a dozen complaints about malfunctioning safety equipment to local authorities in less than four years.
Lawyers are still trying to untangle the events of that tragic night.
In 2008, an Anoka County jury awarded the victims' families a total of $21.6 million after the families' attorneys argued that the deadly collision was caused by a malfunctioning signal. In September, the Minnesota Court of Appeals ordered a new trial, finding the wrong legal standard was used to determine liability.
While the three members of the court panel disagreed on whether the railroad deserved a new trial, they all found Washington County District Judge Ellen Maas did not abuse her discretion when she sanctioned BNSF last year for multiple instances of misconduct.
In her ruling, Maas said the nation's second-largest railroad engaged in a "systematic exploitation of the civil-justice system of a pervasiveness seldom seen outside of John Grisham novels." Maas said BNSF obstructed the state's crash investigation, destroyed evidence and repeatedly misrepresented facts.
Charles Shewmake, BNSF vice president and general counsel who oversees the company's legal department, said in a July interview that BNSF did not interfere with the state's investigation. He agreed that the railroad could have done a better job on the case."I will acknowledge that there are things that happened in the investigation of that case where best practices weren't followed, where in hindsight we could have done some things differently,'' Shewmake said. "In light of the conduct of the trial, we now know that some of this evidence that we didn't think was relevant, was relevant. And certainly if we were not fully cooperative with law enforcement in that case or any other, going forward it certainly would be my intention that we would be."
'Those poor families'
It was about 10 p.m. on Blaine High School's Homecoming night, and three young men a year past high school were heading to a party for a friend about to be deployed overseas to Afghanistan. After watching a high school football game, they took off from Blaine in Brian Frazier's black Chevrolet Cavalier Z24, a sporty sedan with a personalized license plate reading "THE FRAZ."
Frazier and his buddies -- Corey Chase and Harry Rhoades -- had agreed to pick up two girls. The first, Bridgette Shannon, was at her parents' house in Ramsey. To pick up the second girl, the group would have to cross the railroad tracks at Ferry Street, a busy intersection near the Anoka County Fairgrounds that sees about 15,000 cars a day.
They never made it.
At 10:09 p.m. on Sept. 26, 2003, Frazier's car was torn in half by a BNSF train moving about 60 miles per hour. The collision knocked the three unbelted passengers out of the car, with one victim flying more than 170 feet from the point of impact. The steering wheel and a deployed airbag were found next to one of them.
Police couldn't figure out who was driving. The only victim still in the car was found belted into the front passenger seat. All four occupants died immediately from blunt force trauma, according to the Anoka County medical examiner.
It didn't take long for news of the horrific accident to get out. Mike Frazier, Brian's father, remembers seeing a television report on the crash shortly before the end of the 10 p.m. broadcast that night.
"I remember thinking, 'Those poor families,'" he said.
Frazier and the other parents didn't find out their children were dead for at least seven hours.
In its defense, BNSF argued the driver of the car failed to stop at a clearly marked intersection. The company maintained the horn sounded, lights flashed and two crossing gates came down to block traffic. The driver's mistake, BNSF said, was steering around the gate in an attempt to beat the train.
That's not how jurors saw it. They put 90 percent of the blame for the crash on BNSF. The jury concluded the crossing signals failed to operate, according to BNSF's legal filing with the Minnesota Court of Appeals.
The crossing operated unusually on the day of the deadly collision, according to the testimony of a BNSF employee. The warning gate is supposed to automatically drop down about 30 seconds before the arrival of a train, giving drivers plenty of time to stop. But on Sept. 26, the warning system engaged two minutes early for one train, and missed the timing mark several other times that day, trial records show.
Train speeds also were fluctuating. Instead of coming through the crossing at the 60 mph speed limit, one locomotive crawled through at 9 mph, trial records show. BNSF's warning systems were out of sync, according to a railroad expert hired by the families' lawyers.
A railroad employee testified he was aware of problems at the crossing. Randy During, a BNSF worker who maintained the local signal system, said the Ferry Street crossing generated the most trouble calls in the area. He told jurors there was a problem with the electrical system, which is used to send signals to the devices that warn drivers of oncoming trains.
During testified that repairs were made to the Ferry Street gate in April, May and June of 2003, when the crossing generated nine trouble tickets. And just one day before the fatal accident, a BNSF crew was dispatched to fix a problem on the track less than a mile from the crossing, trial records show.
A tough case to crack
One of the first people to make it to the crash scene was BNSF claims representative Lynn Ross, who had been investigating train collisions for about three years, first for Union Pacific and later for BNSF.
Her job on the Ferry Street accident, she testified, was to "investigate this grade crossing, to the best of my ability. I'd gather all of the facts, and do my due diligence for the railroad and for the individuals."
Ross gathered as much evidence as she could about what happened. She brought a tape recorder, a camera and a measuring wheel.
"I was instructed early in my career, when there is a grade crossing accident with a fatality or serious injury, that we do our investigation in anticipation of litigation,'' Ross testified.
After reaching the crash site, which police were still taping off, Ross walked to the front of the train, where she interviewed the engineer and conductor.
Ross did not record their statements. It wasn't company policy. "We do not record crew statements,'' she said in a deposition.
She testified that she could only find one eyewitness to the crash -- the engineer, Bradley Bellmore. In his interview with Ross, Bellmore blamed the driver of Frazier's car, saying the gates and warning lights at Ferry Street worked that night. He said he was blowing his whistle when he spotted Frazier's car going around the crossing gates at 5 or 10 mph, according to her deposition.
"He saw the driver look up into his eyes and his exact words were, 'His eyes were as big as saucers, as if he couldn't believe he wasn't going to make it around,'" Ross said in her deposition. "It was very traumatic."
State trooper Scott Trautner, a reconstruction expert who led the investigation, testified it was the hardest case of his career. Initially, at least, he and his partners couldn't make the railroad's theory of the crash fit the physical evidence.
There were scuff marks, scratching and gouges in the right-hand lane, plus debris from a smashed taillight. To Trautner and his partners, including Ken Drevnick, the evidence indicated the vehicle was struck in its own lane. That didn't gibe with Bellmore's account of the vehicle going around the crossing arm in the other lane, where there were no crash marks. In an attempt to make sense of the contradictions, Trautner and his partners tried to recreate the accident with a similar vehicle. They couldn't do it.
"You have to drive around, pull it in, back it up, turn it, almost like a parallel parking maneuver," Trautner testified. "So you can't get the car to that position with the stop-arm down, without really jockeying the car around at slow speed."
To make the pieces fit, investigators developed a new theory -- that the marks were left after the vehicle was scooped up by the train's cowcatcher and dumped back on the road. In the State Patrol's final report on the crash, Trautner blamed the accident on driver error, and he expanded on that theory during the 2008 trial, saying new evidence did not change his opinion.
Drevnick disagreed. After reviewing the new evidence, he testified it would have been impossible for the car to go around the crossing arm at around 30 mph, as the Chevy's black box indicated. He testified that the State Patrol reached the wrong conclusion because investigators were missing critical information, including the fact that crossing equipment is subject to recurring failures. Drevnick testified he believed the car was hit in its own lane, with the gate up.
"There was a mistake that we assumed information right away," Drevnick testified.
"And that assumption was that the gate was down?" asked Bob Pottroff, one of the families' lawyers.
"Yes,'' Drevnick replied.
Trautner testified BNSF failed to cooperate by ignoring his repeated requests to inspect the locomotive in daylight and failing to disclose other vital information. One of Trautner's biggest frustrations was his inability to obtain raw data from event recorders at the Ferry Street crossing. The recorders, stored inside a bungalow, keep tabs on how fast trains are moving and whether the warning system is working properly.
For 22 years, this crossing had been Craig Hildebrant's responsibility. An electrician with BNSF's signal department since 1976, Ferry Street was one of 127 crossings he monitored.
Despite his experience, Hildebrant made a critical error the night of the accident, Judge Maas concluded.
Instead of bringing a reliable witness like a police officer with him while he checked the recorders -- a BNSF policy -- Hildebrant entered the bungalow by himself. He looked at the data regarding the crash. According to his testimony, it showed that the crossing gate and other warning devices worked properly that night. Before he left, he downloaded the data onto his laptop. He later copied it onto a disc, according to the testimony of another BNSF employee.
When Trautner asked to see the data in the weeks following the accident, Ross testified, she told him he could view a printout at BNSF's Northtown Yard in Fridley, where a railroad employee would tell him how to "interpret the data."
That wasn't acceptable to Trautner, who wanted an independent expert to tell him what the data meant. "I didn't want the railroad to do it because they were a party to the crash,'' he testified in his deposition.
Trautner wasn't the only investigator who wanted to see the data. So did the lawyers who sued BNSF on behalf of the victims' families.
In 2004, plaintiff attorney Sharon Van Dyck asked BNSF to let her team review the event recorder data and other records relating to the crash. "Please preserve all potentially relevant records," she wrote.
Four years later, as the trial was about to get underway, the families' lawyers were still trying to unscramble the railroad's responses. They said it wasn't easy.
Hildebrant's laptop computer, which contained his original download, was returned to the railroad for recycling within six months of the crash as part of a normal equipment replacement practice, he testified. And the disc was never found. Hildebrant's supervisor testified he gave it to BNSF's claims department. Ross testified that she never got the disc.
Hildebrant made multiple copies of the data and stored it in various places, but he had trouble remembering the details. He testified that he "forgot" visiting the bungalow three days after the crash to make a copy of the data. For two years, he also forgot that he transferred a backup copy of the crash data to BNSF's server in Fort Worth, Texas, when his laptop was recycled in January 2004, according to his testimony.
Hildebrant's final data download wasn't turned over to the families' lawyers until about two weeks before the trial in 2008. The trial was well underway when Larry Farnham, one of the families' experts, made a key discovery.
Farnham, who helped design one of the event recorders, found something wrong with the data when he pored over the records at his hotel on Memorial Day. Records from various days had been mingled together, he later testified. He no longer trusted the crash data because he'd seen "evidence of tampering."
Farnham was stunned. "You're not going to believe what I just found,'' he told Pottroff.
For years, Pottroff and his team of lawyers had complained about the company's reluctance to produce requested evidence. Now, the legal team believed it had proof that the company intentionally covered up what happened on the night of the crash.
A bit of 'magic'
When court resumed after the holiday, Hildebrant took the stand again. He described going to the bungalow in 2005, a month or two before the families' lawyers inspected the event recorders. He testified that he changed the settings on one of the event recorders so the data would make sense to the visitors.
Pottroff asked if Hildebrant had done that in order to "manipulate or corrupt" data regarding the collision.
"I did not," Hildebrant testified.
When asked if anybody could have altered data that Hildebrant stored on the company's server, Hildebrant said, "Not that I am aware of. It's password protected, as far as I know."
Later that day, a BNSF attorney asked Hildebrant how four pages of data got combined with other paperwork from the case.
"Somebody magically shuffled them together. ... I don't believe I did it,'' Hildebrant testified.
Farnham went next. After pointing out various irregularities, Farnham testified it was obvious someone had manipulated some of the data he reviewed. He said he couldn't trust the railroad's crash data.
"My only explanation is they came from a different crossing," Farnham testified, referring to the crossing data.
After excusing the jury, Judge Maas said Hildebrant's integrity as a witness may have been "compromised."
"The jury has to decide whether Mr. Hildebrant is lying through his teeth or telling the truth. ... That's kind of where it boils down,'' she told the lawyers gathered in front of her.
Ultimately, Maas concluded BNSF's handling of the event recorder data amounted to misconduct, and she concluded that Hildebrant had lied under oath. She imposed a total of $4.2 million in sanctions on the railroad.
In court filings, BNSF acknowledged a few mistakes, but the railroad said its "less than perfect evidence preservation practices" don't amount to misconduct. Hildebrant didn't commit perjury, the company said, he was forgetful.
"Hildebrant's inability to recall every download and the location of every backup copy is simply not the stuff of sanctions," BNSF said in a November 2008 court filing.
Moreover, the railroad said it cooperated with state investigators, noting that "fact disputes and disagreements are a part of every trial."
At a hearing in 2009, Tim Thornton -- a Minneapolis attorney leading BNSF's appeal -- called Hildebrant a "rogue" employee who "simply wasn't up to the job." But if anyone suffered from Hildebrant's conduct, the company said, it was BNSF. The proof: the record-setting judgment against the company.
Both sides appealed to the Minnesota Supreme Court, which last month agreed to review the families' petition to overturn the appeals court ruling but denied BNSF's appeal. If the case goes to a second trial, the families' lawyers will have to prove that the accident was at least partly caused by the railroad's failure to comply with federal safety regulations, according to the Court of Appeals.
In its majority opinion, the appeals court noted there was evidence showing BNSF complied with federal rules. The railroad had inspected the crossing system once a month and taken other steps to make sure the tracks and signals were operating properly, as required by the government.
In a dissenting opinion, Judge David Minge said there was evidence showing BNSF committed multiple violations of the federal rules, leading him to suggest that a second jury could ultimately reach the same conclusion as the first.
Among those alleged violations: the near-collision suffered by Pam McArthur-Elliott in 2002. In a written statement to the Star Tribune, BNSF said it didn't send anybody to check on the crossing after local police and crew members on two trains found no problems with the warning system.
Eleven months later, police reported another malfunction at the Ferry Street crossing. BNSF sent a crew to find out what was wrong.
"They didn't fix the problem because they couldn't identify a problem," BNSF said in a written statement.
Seven months after that report, Brian Frazier's car entered the same intersection and all four occupants were killed.
Paul Levy 612-673-4419