Testimony before the
Committee on Transportation and Infrastructure
Subcommittee on Railroads, House of Representatives
Regarding Human Factors in Railroad Accidents
March 6, 1996

Good afternoon Madam Chairwoman and Members of the Subcommittee. I appreciate the opportunity to represent the National Transportation Safety Board (NTSB) before your Subcommittee to discuss human factors in railroad accidents. Before I begin my testimony, I want to introduce Mr. Robert Lauby, Chief of our Railroad Division, and Mr. Barry Sweedler, Director of our Office of Safety Recommendations.

Thus far, 1996 has been a busy year for the NTSB, and a tragic year for the railroad and rail transit industries. Since January 1, 1996, the Safety Board has launched railroad investigators to 17 railroad accidents that resulted in 19 fatalities, 230 injuries, and over $64 million in damages. Included in these 16 accident investigations are 6 runaway trains, 5 collisions, 4 derailments, and 1 grade crossing accident.

RECENT SIGNIFICANT ACCIDENTS

Occurrences since January 1 include tragic accidents on the Washington METRO in Gaithersburg, Maryland; the Burlington Northern Santa Fe Railroad in Cajon Pass, California; New Jersey Transit in Secaucus, New Jersey; and Maryland Transit Administration's MARC Train in Silver Spring, Maryland. These four accidents are major accidents that involved launching full teams of NTSB investigators. Other important accidents involving the Burlington Northern Santa Fe Railroad in St. Paul, Minnesota and the Southern Pacific Railroad in Tennessee Pass, Colorado are being investigated as regional accidents by the Safety Board's regional accident investigators. These accidents all involve important railroad safety issues, and today I would like to discuss four of those issues -- fatigue in train operators; positive train separation; signal calling; and training and communication.

Human performance and its impact on safety have become of utmost concern in all modes of transportation. According to the University of Pennsylvania School of Medicine, approximately 70 percent of accidents that occur in all phases of daily life can be attributed to human error. Likewise, in transportation an employee who is inattentive, whether impaired, distracted, bored, fatigued, or who is inadequately prepared or inadequately trained to do his job, can expect to find him or herself involved in an accident sooner or later.

Today I would like to concentrate on three tragic accidents that involve commuter rail operations. I will use these three accidents -- New Jersey Transit accident in Secaucus, New Jersey; Maryland Transit Administration's MARC accident in Silver Spring, Maryland; and the METRA/schoolbus accident in Fox River Grove, Illinois -- to illustrate some important human performance safety issues. These accidents are still under investigation and conclusions concerning their cause are still being developed. The human factors issues discussed here illustrate some of the human factors problems we are facing in transportation safety, and are not intended to be specific conclusions regarding the causes of these accidents.

Fatigue of Train Operators

The first safety issue I would like to discuss is fatigue of train operators. Although fatigue has yet to be identified as a causal issue in the New Jersey Transit accident, the engineer was working on a split shift.

About 8:40 a.m. Eastern Standard Time, on February 9, 1996, New Jersey Transit Commuter Train 1254, operating eastbound from Waldwick to Hoboken, New Jersey, collided head-on with the lead locomotive of New Jersey Transit Commuter Train 1107. Train 1107 was a westbound train operating between Hoboken, New Jersey and Suffern, New York, and consisted of a diesel locomotive and six passenger cars. The train was configured with the diesel locomotive in front and was staffed by an operating crew of three.

The westbound train was operating on a clear signal and travelling at about 53 miles per hour in an area authorized for 60 miles per hour. The eastbound train left Harmon Cove station, accelerated to 53 miles per hour, reduced speed to 30 miles per hour, then to 19 miles per hour, and impacted the westbound train at about seven miles per hour. Eastbound train 1254 had passed a stop signal and fouled the mainline when the collision took place.

There were over 400 passengers on the two trains. Three fatalities, including the engineers on both trains, and 162 injuries resulted from the collision. One passenger who was riding in the cab car at the leading end of the train 1254 was also killed. Of the 162 injuries, 75 were transported to local hospitals and seven were admitted.

The engineer of train 1254 was working a split shift. He reported to work at 6:00 p.m. Thursday, and operated trains until about 1:00 a.m. Friday. From 1:00 a.m. until 5:40 a.m., when he went back on duty operating trains, he rested, using seats in passenger coaches of the train to sleep. He was scheduled to get off work at 7:30 a.m., but he worked overtime and was assigned to operate one more train. He "dead-headed" to Waldwick, New Jersey and operated the accident train from Waldwick to the collision site. The investigation revealed that this crew had been on this schedule all week as well as the previous week. The engineer had worked this schedule for over three years at the time of the accident.

The engineer's work schedule was in compliance with the Federal Hours of Service statute. The statute allows an engineer to work a 12 hour split shift provided that the engineer is given a continuous rest period of at least 4 hours duration between shifts.

On this part of the system, New Jersey Transit controls trains through the engineer and the signal system. There are no automatic stop features on this part of the system. New Jersey Transit relies on the engineer adhering to the signal aspects to maintain safe train operations.

Our investigative team is looking into a variety of issues in the Secaucus, New Jersey accident including:

Positive Train Separation;
Signal Calling by Train Operators;
Crashworthiness of Passenger Cars;
Operator Fatigue and Rest Facilities; and
Operator Fitness for Duty.

Fatigue and its consequences continue to be a pervasive issue in the transportation industry. The Safety Board has investigated numerous accidents in which the effects of fatigue, circadian factors, and sleep disorders may have contributed to accidents. These accidents have raised serious concerns about the far-reaching effects that poor scheduling of work and rest periods have on the performance of operating personnel in all transportation modes.

The Hours of Service Act, passed by Congress in 1907, allows train engineers to work 12 hour duty periods. Thus, the engineer in the Secaucus accident was well within the allowable duty limits. The Hours of Service Act allows the maximum 12- hour duty period to be interrupted (split) by a rest period of at least 4 hours after which the remainder of the 12-hour duty period may be completed. This allows railroads to have the services of a crew for up to 16 hours at a time. The Board is unaware of any research on the effects of this type of split work schedule on fatigue. Irrespective of the engineer's choice of resting place, the Safety Board believes that a 4-hour rest period is not a substitute for 8-hours of continuous sleep.

As a result of many accident investigations by the Safety Board in which the effects of inadequate rest by crewmembers were directly or indirectly associated with the accident cause, the Safety Board in 1989 issued three Safety Recommendations to the Department of Transportation calling for a coordinated and aggressive Federal program to address the fatigue problem in the transportation industry in all of its facets (I-98-1 through -3). As a result, a number of actions have been initiated in the railroad area and in the other transportation modes, notably the aviation industry. The Brotherhood of Locomotive Engineers and the Association of American Railroads, along with a number of individual railroads, have been working to evaluate the work/rest problems of train crews. The AAR has urged its members to provide educational information on fatigue to their employees.

The Board believes, however, that additional research on scheduling practices and fatigue involvement in railroad operations is warranted.

Positive Train Separation and Signal Calling

Positive Train Separation is a system that backs up the action of the human operator to prevent collisions or overspeed derailments. Calling signals is one method that can be used by train crews to help focus their attention at the task at hand and fight the tendency to become distracted. On Friday, February 16, 1996, at about 5:38 p.m., eastbound MARC Train 286, consisting of three passenger cars and a locomotive, collided nearly head on with westbound Amtrak Train 29, the Capital Limited.

The MARC train was operating in scheduled commuter service between Brunswick, Maryland and Washington's Union Station. It was a "push-pull" train with a locomotive on one end and a cab car on the other, which allowed the train to be operated from either end. The engineer was operating the train from the cab car, or in a "push" operation at the time of the collision.

Amtrak Train 29, the Capital Limited, consisted of two locomotives and 15 cars, travelling between Washington, D.C. and Chicago, Illinois. It departed from Washington's Union Station about 20 minutes late because of a delay in servicing and coupling the locomotives to the train.

At the time of the collision, the Amtrak train was travelling west on track 2. The train had been routed on track 2 to pass a standing CSXT freight train. According to signal records and interviews with the crew of the standing freight train, the Amtrak train was operating on an approach medium signal and just beginning to negotiate a crossover from track 2 to track 1 at the time of the collision.

Signal tests are currently being conducted. However, based on the position of the crossover, the MARC train should have had an approach signal about 2.6 miles before the accident scene and a stop signal 170 feet prior to the point of collision. If the signal system was working properly, the MARC engineer should have received an approach signal just before the Kensington Station. The approach signal indicates that the engineer should not exceed 30 miles per hour in the signal block and be prepared to stop at the next signal. The MARC train passed this approach signal and immediately made a stop at the Kensington Station. According to statements from a MARC train operating in the opposite direction, the operator of MARC Train 286 acknowledged the signal over the radio. The other MARC operator did not remember if the operator indicated that the signal was clear or approach. After a short station stop at Kensington, the operator of train 286 accelerated his train to normal track speed. He did not limit his speed to 30 miles per hour as would have been required by a prior approach signal.

According to the event recorder, the MARC train operator applied emergency brakes about 1100 feet before the collision. At the time he applied emergency brakes, the MARC train was travelling about 63 miles per hour. The train had slowed to about 40 miles per hour at the time of the collision. The Amtrak train was travelling at about 30 miles per hour at the time of the collision. The stopping distance of the MARC train from 63 miles per hour is estimated to be about 2,000 feet.

There were 164 passengers, 13 on-board service personnel, four operating crew, and one mechanical rider on the Amtrak train. The Amtrak engineer and assistant engineer were injured in the accident. There were three crewmembers and about 19 passengers on board the MARC train. All three crew members and eight passengers aboard the MARC train were killed in the accident and resulting fire.

The collision between the Amtrak locomotive and the MARC cab car tore away the front left quadrant of the MARC cab car. At the same time, the locomotive fuel tank on the lead Amtrak locomotive ruptured and likely sprayed fuel in the vicinity of the MARC cab car. All three MARC cars and the MARC locomotive were derailed in the accident. Both Amtrak locomotives and the first eight cars were derailed. The derailed Amtrak equipment consisted of six mail cars, one baggage car, and a transition dorm car. The transition dorm car was the only occupied Amtrak car to derail. Issues being looked into include:

Positive Train Separation;
Signal Calling;
Passenger Car Safety Standards;
Emergency Exit Windows;
Emergency Door Operations;
Flammability of Car Components;
Crashworthiness of Passenger Cars and Locomotive Fuel Tanks;
Signal Placement;
Signal System Reliability; and
Attentiveness of Train Operators.

The urgency of the positive train separation issue has been highlighted over and over in Safety Board investigations since 1969. Recent railroad accidents that took place in Sugar Valley, Georgia; Corona, California; Knox, Indiana; Ledger, Montana; Kelso, Washington; and Thedford, Nebraska; could have all been prevented if a fully developed positive train separation (PTS) system had been in place. And now, after the tragic accidents that have occurred in the last two months, we must add two other accidents to this list.

Based on the factual evidence gathered at the scene, the Safety Board investigators feel that the accidents that occurred on New Jersey Transit in Secaucus, New Jersey, and on the MARC Train in Silver Spring, Maryland, could have been prevented if a fully developed positive train separation control system had been in place.

The Safety Board has long been an advocate of advanced control systems that will provide positive train separation and act as a safety net for human performance failures in the operation of trains. Positive train separation is also on the NTSB's list of "Most Wanted" transportation safety issues. The Safety Board believes that new technology -- specifically an advanced electronic system -- can reduce the number and severity of human performance train operations accidents.

Up until December of 1993, the NTSB was discouraged with the pace at which the railroad industry was developing an advanced train control system that could provide positive train separation. Since that time, however, there have been some important developments.

First, the Association of American Railroads (AAR) and the railroads they represent have entered into serious discussions on the AAR's advanced train control system (ATCS) project. The discussions focused on some important issues that need to be fully addressed if ATCS or any other form of positive train separation system is to be implemented. The participants in these discussions have tried to define the system and identify its features. They have also tried to determine the cost and the return on investment for a fully implemented system.


Secondly, the Union Pacific and the Burlington Northern Santa Fe Railroads are working on a joint project to install an advanced train control system that will provide positive train separation on 750 miles of track in the Pacific Northwest. The system is to use a combination of technologies to achieve positive train separation including transponder and global positioning satellites (GPS) to locate and monitor train movements. The test will be the first "real-world" field demonstration of advanced train control technology since Burlington Northern's ARES (Advanced Railroad Electronics System) project was cancelled in 1992. The NTSB is very encouraged with these developments, and we are anxious to see the actual system in operation.

Following the Kelso, Washington, accident, the Safety Board reiterated Safety Recommendations R-87-16 and R-93-12, made to the FRA on May 19, 1987 and July 29, 1993, respectively. These recommendations form the foundation of the Safety Board's effort to achieve Federal standards that will require the installation and operation of a train control system on mainline tracks that will provide for positive separation of all trains. These recommendations also call for the establishment of a firm timetable that includes at a minimum, dates for implementation of a fully developed advanced train control system, and commitment to a date for having the advanced train control system ready for installation on the general railroad system. The recommendations are classified Open--Acceptable Response, based primarily on FRA's July 1994 report to Congress titled "Railroad Communications and Train Control." In that report, the FRA outlined a preliminary schedule for further development of a positive train separation system.


Also as a result of the Kelso accident, the Safety Board also issued new positive train separation recommendations to the FRA and the AAR. Safety Recommendation R-94-16 asked the AAR to identify and evaluate, in conjunction with the FRA, all of the potential benefits of positive train separation and include them in any cost benefit analysis conducted on positive train separation control systems. AAR's initial response stated that it was not possible to determine any business benefit from the PTS test project being conducted in the Pacific Northwest. The AAR went on to say that:

" ... the Federal Railroad Administration (FRA) agrees that railroads are 'justified in insisting that the PTS debate include a clear focus on safety costs and benefits.' With the proper focus on safety benefits, the Report to Congress correctly evaluated the potential benefits of PTS and concluded that it cannot be economically justified at this time."

Safety Recommendation R-94-14 also asked the FRA, in conjunction with the AAR, to identify and evaluate all of the potential benefits of positive train separation and include them in any cost benefit analysis conducted on positive train separation control systems.

The FRA responded that evaluation of the nonsafety business benefits associated with the UP/BN pilot project would not be feasible. The FRA went on to support the AAR and the railroad's position that business benefits should not be assessed. They also conclude that it is not government's role to "substitute our judgment for the judgment of senior railroad managers regarding matters within their special expertise and responsibilities as corporate officers."

The Safety Board believes that the business benefits associated with PTS are real and need to be included in the cost benefit analysis. If safety is the only criteria for justifying PTS, then the growth of PTS will be very slow. The Safety Board is concerned that lack of understanding of the business benefits of PTS may be used to label PTS control systems as -- rightly or wrongly -- too costly for implementation.

While slow progress is being made, we would hope that the recent rash of tragic accidents will spur the FRA and the industry to make a firm commitment to positive train separation and establish a firm timetable for its implementation. The Safety Board was pleased to see in the Federal Railroad Administration's February 20, 1996 Emergency Order that "... the most effective preventative measure is a highly effective train control system," especially automatic systems. We hope that these words will be translated into action.

Signal Calling

Signal calling is an issue being looked into in two of the recent accidents. Signal calling is an activity that can help keep train crews alert and focused on safe operation of the train. Calling signals over a radio also provides management with the opportunity to monitor compliance with signal calling activity.

The Board believes that requiring signals to be called over the radio might have helped control distractions that might have been present in the New Jersey Transit accident in Secaucus, New Jersey and the Maryland Transit Administration MARC Train accident in Silver Spring, Maryland.

In recommendations made in 1976 and 1984, the Safety Board recommended that engine crews communicate fixed signal aspects to other crew members. The FRA did not act on these recommendations, instead opting to evaluate and improve the use of radios for communication in the railroad industry. Nothing concrete resulted from that project and the recommendation was subsequently closed as unacceptable action.

In its February 20, 1996 Emergency Order, the FRA required engineers on trains exceeding 30 mph and not equipped with cab signals or automatic train control, to call out to other crew members on the train, for acknowledgement, the displayed aspect of restrictive signals that are passed.

Training and Communication

Training and communications are human performance activities that affect, not just transportation businesses, but every business in the United States. Millions of dollars are spent by business and industry to train employees and to also teach them how to communicate better. Good communications is known to make businesses more efficient and more profitable; but, how does it affect transportation safety?

About 7:10 a.m., Central Standard Time on Wednesday, October 25, 1995, an eastbound METRA commuter train struck a Crystal Lake School District Bus at the Algonquin Road highway/rail grade crossing in the community of Fox River Grove, Illinois. The train consisted of one locomotive and 7 bi-level passenger cars operating in a push configuration. The train, according to the onboard event recorder, was travelling at about 59 miles per hour when it struck the left rear of the school bus.

Of the 35 student passengers on board the school bus, 7 were fatally injured, and 24 others received injuries varying from minor to critical. The bus driver also received minor injuries.

The school bus driver told investigators that she stopped on the south side of the railroad tracks. While stopped, she opened the passenger loading door, looked right, then left, then right again. At the point where she stopped the bus, she was able to see the railroad tracks in both directions. She stated that she saw no activity in either direction.

The traffic signal at Northbound Algonquin Road was red. The bus driver proceeded across the tracks and stopped at a point where the rear of the bus extended about 30 inches into the space required for passage of an eastbound commuter train. The bus driver stated that she was not aware that the rear of the school bus was extending into the train's space. She was a substitute driver and this was her first time on this route.

The bus driver said that she did not hear the train whistle or the crossing bells. She did not see the flashing lights of the crossing or the crossing gates lowering. She did not hear the crossing gate striking the upper left side of the bus body or decipher any warnings of the train's approach from the student passengers in the back of the bus. The school bus driver told investigators that she never saw the train. At the time of the collision, she was still awaiting a green signal light that would allow her to make a left turn onto U.S. Route 14.

The investigation has found no mechanical deficiencies with the school bus or railroad equipment. Tests of the locomotive whistle and control cab whistle found both to meet the applicable federal standards. Additionally, the light mounted on the front of the cab control car met applicable federal specifications.

The school bus was equipped with a stereo system that incorporated 8 speakers located along the top side of the bus interior. Additionally, a sound absorbing perforated ceiling liner extended the length of the bus interior.

Post accident testing revealed that a northbound vehicle could, under certain circumstances, have only two seconds of green light before a train physically occupied the crossing. Consequently, the Safety Board asked appropriate state and federal agencies to identify and monitor similar highway rail grade crossings and highway intersections to determine if a sequencing problem exists. The Safety Board is pursuing the following issues in this accident:

Bus driver training;
Bus routing policies;
Operation of the combined signal system;
Design of Highway Intersections located in close proximity to Grade Crossings;
Crashworthiness of the school bus; and
Occupant restraints.

Madam Chairwoman, every year about 4,600 motor vehicles are involved in accidents at grade crossings. These accidents kill about 500 people, and they injure more than 1,800 people annually.

With regard to the Fox River Grove grade crossing accident, there are many human factors safety issues being looked into by the Safety Board. In fact, grade crossing accidents often seem to involve human performance issues. In the Fox River Grove accident, training of the bus driver and communications between the different parties involved in the accident are certainly being looked at as issue areas.

The bus driver involved in the accident was in charge of training for the Crystal Lake School District. Although she had responsibility for training other drivers, the day of the accident she was a substitute and was unfamiliar with this particular run. The regular bus driver knew that there was inadequate room at the far side of the crossing for her bus and, therefore, never moved her bus across the grade crossing unless the light was green. The bus driver involved in the accident thought that she had to cross the tracks and trip a sensor before the signal would turn green and allow her to turn west on U.S. Route 14.

The Safety Board believes that individuals and the organizations they work for both have an obligation to ensure that the employees are provided sufficient information to ensure that they can successfully perform their duties.

Communications broke down between the substitute bus driver and the regular bus driver; between the Chicago Northwestern Railroad and the Illinois Department of Transportation; and between the local Fox River Grove authorities and both the Chicago Northwestern Railroad and the Illinois Department of Transportation. Here are just a few of the things that went wrong.

The regular bus driver was required to brief the substitute bus driver with any special information that she needed to complete the run. The briefing was to include safety information. The briefing never took place.

The Chicago Northwestern and the Illinois Department of Transportation were to coordinate the preemption of the signal at Algonquin Road whenever a train tripped the crossing gates. Neither organization understood how the other organization was providing the preemption signals or using the preemption signals.

The local police department had received several complaints about the operation of the Algonquin Road crossing. The police chief was at the intersection at the time of the accident troubleshooting yet another complaint. In response to those complaints, signal maintenance personnel from both the railroad and the highway had been dispatched to monitor their respective signal's operation. The results of these site inspections had generally been that the signal system was working as designed.

The Illinois Department of Transportation modified the area of the crossing by adding two additional lanes of traffic on Route 14 and shortening the traffic storage area for the light. No changes were made to the crossing even though the storage area had been reduced substantially.

The Fox River Grove accident is a good example of how things can go wrong when all parties involved in a grade crossing do not communicate with each other. The Chicago Northwestern Railroad, the Illinois Department of Transportation, the local police department, and even the grade crossing users, have a part to play in grade crossing safety. The railroads have to be involved, and we believe this is a problem that deserves the full attention and support of the railroad industry.

The local governments need to cooperate and close crossings where there are alternate crossings available and to enhance grade crossing safety through the installation of automatic lights, bells, and gates at more crossings. Grade crossings are the "ultimate human factors problem" in the railroad industry. It will take cooperation and communication between all parties to enhance grade crossing safety and prevent future tragedies like that which occurred at Fox River Grove.

Madam Chairwoman, we appreciate the opportunity to discuss these important issues with the Subcommittee. The Safety Board encourages the Federal Railroad Administration to take the necessary action to address the Hours of Service Act, positive train separation, and grade crossing safety initiatives. If rulemaking be required, we would encourage the Federal Railroad Administration to act quickly to establish a consistent level of safety, which would save Americans from death or injury in future railroad accidents.

That completes my statement, and I will be happy to respond to any questions the Subcommittee may have.

Jim Hall's Speeches