Railroad accident report—Derailment of Southern Pacific Transportation Company freight train on May 12, 1989 and subsequent rupture of Calnev petroleum pipeline on May 25, 1989—San Bernardino, California

by National Transportation Safety Board

including: the new destination of the car, a lading code for the car, any special handling associated with the car, and a tonnage figure. This information was entered into the computer system's car file which contains, in addition to the above information, the physical characteristics of each car on the SP system. The yard clerks understood that the tonnage figure would be updated at a later time when the shipper's bill of lading was received in the billing office. SP's director of clerical operations testified that cars are often moved in service before the shipper's bill of lading information is received and entered into the billing system. He further testified that following the train derailment, "We have changed the system so that regardless of what estimate is put into the release, the computer will go to the car file and automatically update that tonnage to the capacity of the car." According to the director of clerical operations, the maximum tonnage figure will remain in the car file of the computer until the shipper's bill of lading is received and only when the bill of lading indicates a shipper-certified weight will the maximum tonnage figure be adjusted to reflect the shipper-certified weight. If an estimated weight is indicated on the shipper's bill of lading, the maximum tonnage figure will remain in the car file of the computer system until the car has been weighed. The nearest scale to the Mojave Yard was at West Colton.

CONTENTS

EXECUTIVE SUMMARY		vi
INVESTIGATION
Events Preceding the Train Derailment		1
⁠Loading of Hopper Cars		1
⁠Preparing the Shipper’s Bill of Lading		3
⁠Activities of Extra 7551 East		5
⁠Activities of Helper Unit		8
⁠Movement of Extra 7551 East from Oban to Hiland		8
The Train Derailment		9
Emergency Response to Train Derailment		15
Wreckage Clearance/Pipeline Surveillance Operations Following the Train Derailment		17
⁠May 12, 1989		17
⁠May 13, 1989		20
⁠May 14, 1989		25
⁠May 15 and 16, 1989		25
⁠May 17, 18, and 19, 1989		30
Events Preceding the Pipeline Rupture		31
Pipeline Rupture		31
⁠Pipeline Operations on May 25, 1989		31
⁠Witnesses’ Observations		33
Emergency Response to Pipeline Rupture		33
Pipeline Surveillance Operations		35
Injuries		37
Damages		37
⁠Train Derailment		37
⁠Pipeline Rupture		39
⁠Damage to the Pipeline		39
Track and Signal Information		42
⁠Track		42
⁠Signals		43
Train Information		43
⁠Locomotive Units		43
⁠Hopper Cars		44
Mechanical Information		45
⁠Use of Dynamic Brakes		45
⁠Maintenance Reports and Reporting of Defective Locomotive Units		46
⁠Recovering Dynamic Brakes		47
⁠FRA’s Position Regarding Functioning Dynamic Brakes		48
Southern Pacific’s Method of Operation		48
⁠Air Brake Rules and Timetable Instructions		48
⁠Communication Between Head-end and Helper Engineers		53
⁠Tonnage Information for Cars		53
The Calnev Pipeline		54
⁠Description		54
⁠Check Valves		55
⁠Block Valves		56
⁠Dispatch Center		56
⁠Emergency Response Manual		57
Personnel Information		58
⁠Operating Crew of Extra 7551 East		58
⁠Other Southern Pacific Personnel		60
⁠Calnev Pipeline Dispatcher		61
Southern Pacific Training Programs		61
⁠Engineer Training Program		61
⁠Dispatcher Training Program		63
⁠Clerk Training Program		64
Calnev Pipeline Dispatcher Training Program		64
Southern Pacific Management Oversight of Train Operations		65
Industry Pipeline Standards and Federal Regulations		66
Oversight of Calnev Pipeline Operations		67
Meteorological Information		69
Medical and Pathological Information		69
⁠Train Derailment		69
⁠Pipeline Rupture		70
Toxicological Information		70
Southern Pacific’s Physical Examination Policy		71
Tests and Research		71
⁠Event Recorders		71
⁠Train Dynamics Analyzer Runs		72
⁠Instrumented Brake Shoe Tests		73
⁠Train Vibration Study		73
⁠Soil Inspection Report		73
⁠Metallurgical Testing		75
⁠Simulation of Excavating Equipment Operations		79
Other Information		80
⁠Train Movements Following the Train Derailment and Preceding the Pipeline Rupture		80
⁠Agreement Between the Southern Pacific and the City of San Bernardina Following the Train Derailment		80
⁠Development of Land Adjacent to the Southern Pacific Railroad and the Calnev Pipeline		81
⁠Disaster Preparedness		82
⁠Train Derailments over Pipelines		82
ANALYSIS
General		85
The Train Derailment		85
⁠Axles of Dynamic Brakes		86
⁠Trailing Tonnage		87
⁠Operation of Extra 7551 East Down the 2.2 Percent Grade		90
⁠Derailment Speed		92
Communication		92
Testing Dynamic Brakes		94
Event Recorders		95
Computer-Generated Tonnage Profile Information		97
Dynamic Brake/Emergency Interlock		98
Reporting Defective Conditions on Locomotives		98
Training Program for Engineers		99
Southern Pacific Training Program for Yard Clerks		99
Southern Pacific Management Oversight of Train Operations		100
The Pipeline Rupture		101
⁠Removal of the Train Wreckage		103
⁠Removal-of Trona From Over the Pipeline		103
⁠Excavation and Inspection of the Pipeline		104
⁠Removal of Trona From the Oer ailment Area		105
Adequacy of Calnev's Inspection cf the Pipeline Following the Train Derailment		105
The Timing of the Pipeline Rupture		108
Calnev Pipeline Monitoring System		108
Shutdown of Failed Pipeline		109
⁠Check Valves		109
⁠Remotely Operated Valves		110
⁠Federal Reguations		112
Enhancing Public Safety Near Railroads and Pipelines		115
Survival Aspects		117
Emergency Response		117
Medical and Toxicological Factors		119
Southern Pacific's Physical Examination Policy		119
CONCLUSIDNS
Findings		120
Probable Cause		123
RECOMMENDATIONS		124
APPENDIXES
Appendix A—Investigation and Hearing		129
Appendix B—Personnel information		130
Appendix C—Bill of Lading Information		131
Appendix D—Tonnage Profile of Extra 7551 East		137
Appendix E—OPS Hazardous Facility Order and Subsequent Amended Orders		144
Appendix F—Assessment of Damages to Residences and Property		153
Appendix G—FRA Letter Regarding Functioning Dynamic Brakes		155
Appendix H—Southern Pacific Timetable Instructions (Maximum Tons per Operative Brake)		156
Appendix I—Selected Provisions of ASA Code B31.4		157
Appendix J—Pertinent Provisions of 49 CFR Part 195		159
Appendix K—Pertinent Provisions of 49 CFR Part 192		165
Appendix L—Alert Bulletin Issued by RSPA on November 13, 1989		166
Appendix M—Strip charts From Event Recorders of Extra 7551 East		171
Appendix N—Report of Converse Consultants		174
Appendix O—Agreement Between the Southern Pacific and the City of San Bernardino		183

Executive Summary

About 7:36 a.m., Pacific daylight time, on May 12, 1989, Southern Pacific Transportation Company freight train 1-MJLBP-111, which consisted of a four-unit locomotive on the head end of the train, 69 hopper cars loaded with trona, and a two-unit helper locomotive on the rear of the train, derailed at milepost 486.8, in San Bernardino, California. The entire train was destroyed as a result of the derailment. Seven homes located in the adjacent neighborhood were totally destroyed and four others were extensively damaged. Of the five crewmembers onboard the train, two on the head end of the train were killed, one received serious injuries, and the two on the rear end of the train received minor injuries. Of eight residents in their homes at the time of the accident, two were killed and one received serious injuries as a result of being trapped under debris for 15 hours. Local officials evacuated homes in the surrounding area because of a concern that a 14-inch pipeline owned by the Calnev Pipe Line Company, which was transporting gasoline and was located under the wreckage, may have been damaged during the accident sequence or was susceptible to being damaged during wreckage clearing operations. Residents were allowed to return to their homes within 24 hours of the derailment.

About 8:05 a.m., on May 25, 1989, 13 days after the train derailment, the 14-inch pipeline ruptured at the site of the derailment, released its product, and ignited. As a result of the release and ignition of gasoline, 2 residents were killed, 3 received serious injuries, and 16 reported minor injuries. Eleven homes in the adjacent neighborhood were destroyed, 3 received moderate fire and smoke damage, and 3 received smoke damage only. In addition, 21 motor vehicles were destroyed. Residents within a four-block radius were evacuated by local officials.

Total damages as a result of the train derailment and pipeline rupture exceeded $14 million.

The major safety issues include:

Railroad

• the means by which the shipping weights were determined for the shipment of the trona laden hopper cars;
• the dispatching of locomotives without operable dynamic brakes on mountain gradients;
• the information received by the road engineer regarding the weight of the train and the number of operable dynamic brakes;
• the communication between the road and helper engineers regarding the operation of the train, and communication with the dispatcher;
• the engineer’s training program, which did not adequately address emergency situations;
• changes in operating procedures made by Southern Pacific after the accident;

Pipeline

• Southern Pacific’s wreckage clearing operations in the area Calnev’s pipeline alignment;
• Calnev’s oversight surveillance of the train wreckage clearing operations and trona removal in the derailment area;
• Calnev’s assessment of pipeline integrity prior to resuming full pressure operation of the pipeline after the derailment;
• the effectiveness of the pipeline check valves used to minimize product release;
• the adequacy of Federal regulations to address the inspection and maintenance of valves for liquid pipelines.

The National Transportation Safety Board determined that the probable cause of the train derailment on May 12, 1989, was the failure to determine and communicate the accurate trailing weight of the train, failure to communicate the status of the train’s dynamic brakes, and the Southern Pacific operating rule that provided inadequate direction to the head-end engineer on the allowable speed and brake pipe reduction down the 2.2-percent grade.

The National Transportation Safety Board determined that the probable cause of the pipeline rupture on May 25, 1989, was the inadequate testing and inspection of the pipeline following the derailment that failed to detect damage to the pipe by earth-moving equipment. Contributing to the cause of the pipeline rupture was the severity of the train derailment that resulted in extensive wreckage and commodity removal operations. Contributing to the severity of the damage resulting from substantial product release was Calnev’s failure to inspect and test check valves to determine that they functioned properly, particularly after the train derailment.

As a result of its investigation, the Safety Board issued safety recommendations to the Southern Pacific Transportation Company, the Calnev Pipe Line Company, the Federal Railroad Administration, the Association of American Railroads, the City of San Bernardino, the Research and Special Programs Administration, the National Association of Counties, and the National League of Cities. The safety board also reiterated safety recommendations to the Research and Special Programs Administration and the Federal Railroad Administration.

National Transportation Safety Board

Washington D.C. 20594

Railroad Accident Report

Derailment of Southern Pacific Transportation Company
freight train on May 12, 1989, and subsequent
rupture of Calnev Petroleum Pipeline on May 25, 1989
at San Bernadino, California

Investigation

Events Preceding the Train Derailment

Loading of Hopper Cars.—The Lake Minerals Corporation, an Owens Lake, California, company involved in the mining and shipment of trona,^[1] contracted with the Southern Pacific Transportation Company (SP) to have a shipment of trona transported from the Corporation’s rail facility in Rosamond, California (see figure 1), to the Port of Los Angeles. The trona was then to be loaded into a vessel destined for Colombia, South America. Lake Minerals’ customer had ordered 6,835 tons of trona. The contract with the SP specified that the railroad would provide 69 100-ton open-top hopper cars; Lake Minerals’ payment to the SP was to be based on 100 tons per car.

Because Lake Minerals Corporation did not have rail facilities at its Owens Lake plant, the trona was shipped by truck from there to the rail facility at Rosamond, where the trona was loaded into the open-top hopper cars by a loading contractor hired by the Lake Minerals Corporation. The Lake Minerals Corporation had shipped trona by rail to the Port of Los Angeles on only one previous occasion. The superintendent of Lake Minerals Corporation testified that on that first shipment the company had averaged 88 tons per car when the contract had also called for 100 tons per car. He stated, "We ended up with a significant shortage at the port and did not have enough material to fill the vessel," and "…we ended up with a dead freight charge." For the second shipment, Lake Minerals Corporation requested that the loading contractor install a sensing device on the front-end loader to measure the amount of material that was being loaded into the cars. To test the accuracy of the sensing device, a truck was loaded with the trona and weighed on the truck scale at the loading facility. The device was checked for accuracy after about half the cars had been loaded. The superintendent stated that he was satisfied that the device accurately weighed the loads. He further testified that "we were very concerned with being as accurate as possible." In addition to expressing concern that they did not underestimate the amount of trona loaded, he stated, "At the Port facility there is no way to handle the trona if we had excess material and the vessel was loaded. We would have had to dump it on the ground and haul it back, and we wanted to avoid that at all costs."

Figure 1.—Mojave Subdivision.

Because the rail facility at Rosamond would not accommodate 69 cars, on May 5, May 6, and May 8, 1989, the SP moved 32, 15, and 22 loaded cars, respectively, from Rosamond to a side track at Fleta (figure 1). After the cars were loaded, yard clerks at Mojave "released" the cars by changing the status of each car from an "empty" to a "load,"^[2] in SP's computer system. The computer process required, at the time the status was changed, the entry of an estimated weight of the product. Three different yard clerks, based on their prior railroad experience, entered estimated weights into the car file^[3] of the computer system on three separate occasions—each time the groups of cars were moved from the Rosamond facility to the side track at Fleta. (The 32 cars moved on May 5 were estimated at 50 tons each, the 15 cars moved on May 6 were estimated at 75 tons each, and the 22 cars moved on May 8 were estimated at 60 tons each.) The light (empty) weight of the car was programmed into the system, and the system would automatically compute the total weight of each car. According to their testimony, the yard clerks, who had no knowledge of the contents of the contract between the SP and Lake Minerals, believed that the weight they estimated when the cars were released would be automatically replaced in the computer system by the weights shown on the shipper’s bill of lading when that document was later received in Los Angeles and the shipper weights were entered into the computer. Testimony by the yard clerks further indicated that estimated weights supplied when cars were released were routinely overridden by shipper weights at later dates, and that they had no reason to believe that it would not be done in this instance. One of the yard clerks, who had worked in that capacity for 17 years with the SP and who estimated the weights of the 15 cars moved on May 6, stated that it was important to estimate as closely as possible the actual weights of the cars; however, he could not offer a precise reason for why it was important. There was no documentation available to the yard clerks that indicated the actual weight of trona (or any other commodity).

Preparing the Shipper’s Bill of Lading.—On May 6, 1989, the superintendent of Lake Minerals Corporation submitted a bill of lading for the 69 cars loaded with trona to a shipping clerk at the SP’s yard office at Mojave. The bill of lading (appendix C) indicated the total number of cars to be shipped, the destination of the cars, and the car numbers. The weight of the cars was not listed on the bill of lading, and there was no discussion regarding the weight of the cars. The document was reviewed and signed by both the shipping clerk and the superintendent. The superintendent testified that it was an oversight that he did not provide the weights on the bill of lading. He stated, "There was no question about the weights and it was understood, as far as I knew, that they were 100 ton cars, they were loaded and we’d ordered 69 of them." The shipping clerk testified that after the superintendent of Lake Minerals Corporation left the office, he realized that the SP billing office in Los Angeles would require that a weight be shown on the shipper’s bill of lading. He stated that he attempted to contact Lake Minerals Corporation to inquire about the weights of the cars but was unable to obtain the company’s telephone number. Based on his experience working for the railroad, he then estimated the weight of the product to be 60 tons per car and wrote the figure of 120,000 pounds per car on the bill of lading (appendix C). He testified, "…I figured these cars were lighter than cement cars and I knew cement cars were 75 tons, so my estimated weight was 60 tons and I entered it." The shipping clerk did not indicate on the bill of lading that the weight listed was an estimated weight. After the writing the figure of 120,000 pounds per car on the bill of lading, he sent the document, via a facsimile (fax) machine, to the Los Angeles office. The shipping clerk testified that he had never before received a bill of lading that did not have the weights provided. There was no documentation available to the shipping clerk that indicated the actual weight of trona (or any other commodity) or outlined the procedures to follow when the shipper did not provide weights on the bill of lading. The superintendent of Lake Minerals testified that he believed the weight of 200,000 pounds per car had been written on the bill of lading for the first shipment of trona.

Upon receipt of the document in the Los Angeles office, a billing clerk entered the bill of lading information into SP’s computer system; information that would later be used to prepare the train (tonnage) profile.^[4] According to SP’s director of system clerical operations, there are two methods available to the billing clerk to enter bill of lading information into the computer when a unit train^[5] is involved. He testified, "One is where the only thing that you show is the total shipment weight, the cumulative weight of all cars and not the individual weights of each car. The second method of entry is where you make the individual weights for the individual cars." Further testimony indicated that if the first method is used, weight information will be entered into the waybill file but that any weight previously entered into the car file will not be upgraded. If the second method is used, the weights estimated and previously entered into the car file of the computer system by the yard clerks would be overridden by the weights entered by the billing clerk. The billing clerk in Los Angeles on May 6, 1989, used the first method for entering the bill of lading information. There was no indication on the document received by the billing clerk in Los Angeles that the figure of 120,000 pounds per car was an estimated weight.

Activities of Extra 7551 East.—At 5:00 p.m., on May 11, 1989, the chief train dispatcher on duty at Los Angeles, California, telephoned a yard clerk at Mojave (see figure 1) and informed him of plans to operate a train to haul the 69 cars of trona from Fleta to West Colton, near Los Angeles. At 9:00 p.m. that evening, a traincrew consisting of a locomotive engineer, a conductor, and a brakeman reported for duty at SP’s yard office in Bakersfield, California. At 9:15 p.m. while in the Bakersfield yard office, the conductor telephoned the yard clerk at Mojave and was told about the crew’s assignment to operate SP train MJLBP1-11 (designated Extra 7551 East) out of Mojave to haul 69 cars of trona. The crewmembers were transported in a company van from Bakersfield to Mojave where they arrived and entered the yard office at about 10:30 p.m. The crew picked up a clearance form, train orders, train list, and tonnage profile (the latter document is generated by the SP computer system and based, in part, on information in the car file) (appendix D), and departed the office. The documents provided to the crew indicated that the train consisted of 69 loaded cars with a trailing tonnage of 6,151 tons. The engineer testified that neither he nor the conductor had any concern about the paperwork received. The dispatcher on duty at 5:00 p.m. that day had arranged for the crew to take three locomotive units from the Mojave yard to Fleta (3 miles away) where they would couple onto the 69 cars assembled in the siding. They were to then pick up an additional locomotive unit at Palmdale Two (figure 1) to help in ascending the 2.2 percent grade to Hiland.

After departing the office, the crew proceeded to the yard to check out the three-unit locomotive consist. Between 11:00 p.m. and 11:30 p.m., the conductor called the yard clerk and informed him that locomotive unit SP 7551 was "dead-in-consist" and could not be started. The engineer testified that the crew attempted to determine the reason the unit would not start but was unsuccessful. The yard clerk instructed the crew to use another unit (SP 8278) that was in the yard next to the three-unit consist. The yard clerk then informed the assistant chief dispatcher, who had come on duty in Los Angeles at 10:30 p.m., of the condition of SP 7551 and of the use of SP 8278. The assistant chief dispatcher testified that he was concerned that with only three locomotive units the train could not take the 69 loaded hopper cars farther than Denis (see figure 1), and so he decided to alter the plan to supply locomotive power for Extra 7551 East that had been arranged by the dispatcher on the previous shift. Rather than have the crew pick up an additional locomotive unit at Palmdale Two, the assistant chief dispatcher arranged for a helper locomotive to move toward Mojave, meet Extra 7551 East at Oban, and assist the train up the ascending grade to Hiland and through the Cajon Pass.^[6] The assistant chief dispatcher testified that he made this decision based on his belief that the tonnage of Extra 7551 East was about 8,900 tons, a figure that he calculated based on his experience with the product. He further testified that even though he had a copy of a yard list prepared by the yard clerks the previous week when they released the cars indicating a trailing tonnage of 6,151 tons, he believed that figure to be an estimated weight that would have been overridden when the bill of lading information was placed in the computer system. According to his testimony, he determined the number of locomotive units that would be needed to move the train up the grade based on the 8,900 tons. He testified also that he had never previously recalculated the tonnage of a train to determine the number of locomotives that would be needed. He stated that he further believed that the crew had been provided with an upgraded weight reflecting the figure of 8,900 tons. He did not communicate with the crew nor did he use the computer system, which was available to him at his desk in Los Angeles, to determine the tonnage figure that had been provided to the crew.

After conducting an initial terminal air brake test,^[7] the crew of Extra 7551 East departed Mojave yard (MP 381.3) at 12:15 a.m., on May 12, en route to Fleta (MP 384.4) with a locomotive consisting of units SP 8278, SP 7551, SP 7549, and SP 9340 configured in that order from east to west. (The engineer testified that because he was not told to do anything with unit SP 7551, he kept it in the consist.) The engineer was operating from the lead unit, SP 8278, en route to Fleta.^[8] Because maintenance-of-way equipment was occupying the east end of the siding at Fleta, the dispatcher instructed the crew to continue eastward to Ansel (MP 390.4) and enter a side track at that location to clear the main track for traffic. According to the engineer, Extra 7551 East arrived at Ansel at 12:40 a.m., waited for the main track traffic to pass, and departed Ansel at 1:15 a.m. to return to Fleta. On the return trip to Fleta, SP 9340 was the lead unit in the consist, and the engineer operated from that unit. Because the maintenance-of-way equipment was still occupying the east end of the siding at Fleta, the crew was unable to position their locomotive units on the east end of the train to continue their eastbound trip. It was necessary, therefore, for the crew to enter the west end of the siding (see Figure 2), couple their units to that end of the 69 hopper cars, return westbound to Mojave yard, reposition their locomotives units at that location, and then continue their eastbound train movement. The engineer testified that before departing Fleta, the train line pressure was charged but an air brake test was not conducted. The engineer stated that while operating from unit SP 9340 on the return trip to Mojave, the dynamic brakes^[9] were intermittent: "It would load and then the dynamics would drop out on the unit." (Additional discussion occurs under Mechanical Information.) The engineer testified that after the locomotive consist was repositioned and coupled to the cars in Mojave yard, a test for leakage of the train line pressure and an initial terminal air brake test were performed. According to the engineer, none of the crewmembers expressed concern about the tests. After waiting for an inbound train to clear the main track, Extra 7551 East departed Mojave at about 3:35 a.m. with the engineer operating the train from the lead unit, SP 8278. The conductor was

seated in the cab across from the engineer; the brakeman was seated in the cab of the third unit, SP 7549. According to the engineer, the brakeman was seated in the third unit to keep warm because the second unit, SP 7551, was not operating. The engineer stated that the dynamic brakes on SP 8278 were "working," and that when he asked the brakeman about the condition of the dynamic brakes on SP 7549, the brakeman replied, "It's revving." The engineer further stated that he did not conduct a visual observation^[10] of SP 7549 to determine if its dynamic brakes were operative. Extra 7551 East proceeded to Oban, and the dispatcher instructed the crew to move into the siding at that location to await a westbound train that was being assisted by a helper unit; the helper unit would be cut off and used to assist Extra 7551 East over the Cajon Pass.

Activities of Helper Unit.—At 1:30 a.m., on May 12, 1989, an SP crew, consisting of a locomotive engineer and brakeman, reported for duty at West Colton yard. The crew was transported in a company van from West Colton yard to Dike (MP 481) (see figure 1), arriving at that location at about 2:30 a.m. The crew took charge of a two-unit locomotive consist, SP 7443 (facing west) and SP 8317 (facing east), that was to be used in helper service (assisting trains traversing Cajon Pass). The crew (hereinafter referred to as the helper engineer and the helper brakeman) was instructed by the train dispatcher to operate from Dike to Palmdale Two (MP 417.3) and then to assist a westbound train, Extra 8240 West, between Palmdale Two and Oban (MP 399.9). The helper engineer had been informed by the engineer whom he had relieved that the dynamic brakes on unit SP 8317 were inoperative. The movement from Palmdale Two to Oban was uneventful, and the crewmembers had no concern about the operation of the train. At about 5:06 a.m., the dispatcher instructed the helper engineer to couple the helper locomotive onto the rear of an eastbound train, Extra 7551 East, that was waiting in a siding at that location for helper service through the Cajon Pass.

The helper engineer testified that he did not receive any information from either the head-end engineer or the dispatcher regarding the tonnage of Extra 7551 East nor did he request that information. There was no SP requirement that he be furnished that information. He stated that he did not normally operate over this territory and, therefore, did not know if it was customary to receive that information. He stated further that for the territory over which he normally operated, he usually received that information, and that if he did not, he would request it.

Movement of Extra 7551 East From Oban to Hiland.—After the helper engineer radioed the head-end engineer and informed him that the helper locomotive was coupled onto the rear of Extra 7551 East, an airbrake test was performed; neither engineer noted any deficiencies in the operation of the brakes during the test. Upon receiving a clear signal, Extra 7551 East departed the siding at Oban. At about 5:30 a.m., the helper engineer informed the head-end engineer, by radio, that the trailing units had cleared the siding. The helper engineer testified that his locomotive was in the eighth throttle notch (full throttle) before entering onto the mainline. The head-end engineer and the conductor were still on the lead unit, SP 8278, and the head-end brakeman remained on the third unit, SP 7549. The helper engineer and the helper brakeman were located in the trailing unit, SP 7443, of the helper consist. The helper engineer stated that the trip from Oban to Hiland (MP 463) was uneventful.

The Train Derailment

Testimony indicates that there was no communication between the head-end engineer and the helper engineer from the time Extra 7551 East left the siding at Oban until about 7:03 a.m. when the head-end engineer was cresting the hill at Hiland. The head-end engineer stated that he crested the hill at Hiland (MP 463) at 25 mph or 5 mph below the speed he believed was allowed based on the information he had about the train—6,151 trailing tonnage and four units (two head-end units and the two helper units) with full dynamic brakes and one head-end unit with intermittent dynamic brakes. As he crested the hill, the head-end engineer began using his dynamic brakes and initiated a 6-lb reduction of the air brake pipe pressure. He then asked the helper engineer if he had "…all of your dynamics…." The helper engineer responded, "Yeah, I'm in full." The head-end engineer testified that based on the helper engineer's response he believed that both helper units had operative dynamic brakes and had no reason to believe otherwise. He had not been informed by either the dispatcher or helper engineer that one of the helper units had inoperative dynamic brakes, and he did not inquire about the condition of the dynamic brakes on the trailing units. The helper engineer stated that he did not believe it was necessary for him to alert the head-end engineer of the status of the dynamic brakes on the helper unit because he (the helper engineer) believed the dispatcher would have already made that information known to the head-end engineer. The assistant chief dispatcher, who arranged for the helper unit to assist Extra 7551 East, testified, "I think the normal procedure would be for the helper engineer to relay that information to the road engineer, certainly not the train dispatcher." SP had no requirement that the dispatcher record or disseminate this information.

As the train continued descending the hill, the speed of the train increased to about 30 mph and the head-end engineer increased the brake pipe pressure reduction to 10 psi. According to the head-end engineer, the speed of the train held at 30 mph for a short time and then began to increase. He then increased the brake pipe pressure reduction to about 14 psi. He continued to increase the brake pipe pressure reduction gradually. Each time he reduced the brake pipe pressure, the train's speed would slow slightly and then it would begin to increase again. By the time he reached Canyon, he had reduced the brake pipe pressure a total of 18 psi, but the train was traveling at a speed of 31 mph and accelerating. The head-end engineer stated to Safety Board investigators, "As you're coming down Canyon [MP 469], there are a few places there where it [the train] will run on you, meaning that it's less curvy…you no longer had that resistance of the curves so the train will pick up a little speed, but I was compensating fine." As the train entered straight track, around MP 447, the speed of the train increased, and the engineer began increasing the brake pipe pressure reduction. He stated, "I kept waiting for it [the train] to settle down….I was already up to 20 pounds. Now I knew that was probably enough when that train should start bogging [slowing] down." According to the head-end engineer, he then went into a full service reduction (26 psi). He stated further, "When I made a full service and it wasn't slowing down, we realised that…this train wasn't going to stop." About 7:30 a.m., based on the readout of the event recorder, as the train speed reached 45 mph, the helper engineer, without communicating with the head-end engineer, placed the train brakes in emergency. According to the helper engineer, he did not communicate to the head-end engineer that he was going to place the train brakes in emergency because "at that point there might have been something wrong up there and the speed we were going, corrective action had to be taken and soon…" He further stated that he did not believe that communication prior to that time was necessary because by observing the brake pipe gauge on the rear end, he could tell that the head-end engineer was attempting to take corrective action. According to the head-end engineer, after the helper engineer placed the train brakes into emergency. he placed his brake valve in emergency and the train then began to "surge." According to SP, its locomotives are designed so that when the train brakes are placed in emergency, the dynamic brakes are pneumatically blocked out; both engineers testified that they were aware of this feature. The head-end engineer stated that when the train brakes were placed in emergency he believed there were no longer any options available for controlling the speed of the train.

A motorist who routinely travels on a highway that parallels the railroad tracks for some distance and normally sees trains at that time of the morning testified that she observed "…one train…going a lot faster than some I had normally seen before." The motorist, who estimated that the highway was about ¼ to ½ mile from the tracks, also testified that the train was engulfed in what she assumed to be smoke, which she described as light blue in color. The helper brakeman testified that after the helper engineer placed the brakes in emergency. he observed smoke coming from underneath the train. The head-end engineer also testified that when he looked back over his train, he saw a "lot of smoke coming from the train."

The speed of Extra 7551 East continued to increase as the train descended the hill. The head-end engineer stated that when he realized the train was not slowing, he instructed the conductor to "get on the phone and tell them we got a runaway train." According to a transcript of the dispatcher's radio log, at 7:33:21, an attempt was made to contact the Saugus dispatcher but was not successful. At 7:33:48, the conductor contacted the assistant general yard master at West Colton and informed him, "We have a slight problem. I don't know if we can get this train stopped. We're coming out of Dike [MP 481]." The helper engineer testified that when he overheard the radio transmission to the West Colton yard, he did not believe that the message conveyed the seriousness of the problem and that "I got on there and I called Mayday Mayday to clear the radio waves." He further stated that because the train speed was rapidly increasing, he positioned himself on the floor behind his control stand with his back and head braced against the back panel and his feet braced against the control stand. He stated that he had the radio in his hand, was calling out the speeds and was attempting to call somebody, and that he remembers "calling out the speed when we hit ninety." The helper brakeman stated that he remained in his seat. The transcript of the dispatcher's radio log indicates that at 7:37:09 the following message was transmitted: "Mayday! Mayday! 7551, West Colton-AGYM [assistant general yard master], we're doing 90 miles per hour' nine zero, out of control, won't be able to stop till we hit Colton." The head-end engineer stated that after the conductor called West Colton, "there was nothing left to do." He further stated that he and the conductor remained in their seats and that he believed the speed of the train reached 100 mph. He stated, "The speedometer only went up to 80, but it was way past that….It was as far as it could go."

As Extra 7551 East approached MP 486.6 and entered a 4-degree right-hand curve, the entire train derailed to the outside of the curve; many of the cars crashed into a neighborhood of houses adjacent to the railroad right-of-way (figures 3 and 4).

The dispatcher's radio log indicated that a call from Extra 7551 East stating that the whole train was on the ground was received at 7:37:55. The helper engineer testified that he made the radio transmission after the derailment and that because he had received no communication from the head end, he instructed the helper brakeman to go to the front of the train.

Shortly after 7:30 a.m., two San Bernardino police detectives, who were traveling westbound on Highland Avenue approaching California Street, observed what they stated appeared to be a large flash of light and a large cloud of dust come from the area af Highland Avenue and west of Macy Street. They continued westbound on Highland Avenue, and as they drove past Macy Street, they observed that an SP train had derailed and had crashed into several houses on Duffy Street. One of the detectives used his police radio to advise his dispatcher of the situation and to request emergency personnel. They parked their vehicle on the north side of Highland Avenue and ran up the railroad levee^[11] to evaluate the damage. Several other people had also stopped their vehicles and ran up the levee.

A Southern California Gas Company employee stated that he and another gas company employee were about 100 yards west of Highland Avenue when they observed the train derail at a high rate of speed. He further stated that he immediately ran to the site of the derailment and, along with other unidentified people, helped the engineer who was attempting to pull himself out of the lead locomotive unit. According to the gas company employee, the engineer began looking for his "partner" (who was later identified as the conductor) whom he found fatally injured in the same locomotive unit. After they helped to lay the engineer next to a fence in the rear yard of 2304 Duffy Street to await the arrival of emergency personnel, the gas company employees began shoveling dirt around one of the locomotives to prevent the spilled diesel fuel from spreading. They then began shutting

Figure 3.—Aerial view of train derailment.

Figure 4.—Wreckage distribution and location of Calnev excavations.

off gas lines to the houses that were damaged in the derailment. According to one of the gas company employees, there were no fires associated with the spilled fuel oil or the broken gas lines.

Emergency Response to Train Derailment

The San Bernardino County’s 911 emergency number was called about 7:41 a.m. by a resident who reported that a train was off the tracks and into some houses.

The San Bernardino battalion chief’s unit was the first fire department unit to arrive at the derailment side about 7:48 a.m. The battalion chief stated that in addition to observing the derailed freight train and damaged houses, he noted that a white powdery substance that had been dumped by the train when it derailed was piled over the entire wreckage site. He stated further that he requested a hazardous materials unit to respond to the scene because of the unknown product being carried by the train, the leaking diesel fuel from the overturned locomotives—even though there was no evidence of fire—and the possibility of pipeline involvement.^[12] The battalion chief stated that he was aware that a pipeline was in the area of the derailment but was uncertain of its location at that time.

Police units began arriving also about 7:48 a.m. and began setting up road blocks, evacuating occupied houses, and handling crowd control. An estimated 63 persons were ultimately evacuated from 27 houses in the immediate area of the derailment. As other fire companies arrived, they were placed in strategic locations around the accident site. About 7:55 a.m., fire department personnel began a house-to-house search for survivors. About 11 houses had been impacted by the derailing train. At that time, a canvass of the neighborhood and residents found that no one was reported as missing. About 8:01 a.m., however, a parent reported that two children who resided at 2348 Duffy Street were missing. A second search began and about 8:25 a.m., the first child was found dead; about 10:15 a.m., the second child was also found dead.

Meanwhile, about 8:05 a.m., the San Bernardino deputy fire chief arrived on scene, was advised of the situation by the battalion chief, and then assumed control of the emergency as incident commander. He stated that he approached representatives of Calnev and SP, who had arrived on scene between 8:30 a.m. and 9:00 a.m., and informed them that he was the incident commander in charge. He stated further that by the time he had arrived, the city’s joint response and mutual aid plan had been implemented as a result of the battalion chief’s initial request for additional assistance. The incident commander subsequently established a command post at the corner of Donald and Duffy Streets. The deputy fire chief testified that all subsequent actions by Calnev and SP were coordinated with him. He further testified that because the product that was scattered over the derailment side had been transported in open top hopper cars, he did not believe it was a "serious hazardous material." He was informed initially by SP personnel that the product was potash; later in the day he received a data sheet from the Office of Emergency Services (OES) that identified the product as sodium carbonate.

About 10:40 a.m., the search team was notified that a third person was reported missing at 2326 Duffy Street. Because of the total destruction of the house and the unstable condition of the train cars that were piled up in the area, search and rescue efforts for the missing person at that location were delayed until heavy equipment could be brought in to move some of the damaged structure and train cars.

Representatives from the California OES, which was notified of the accident at 7:45 a.m., through the San Bernardino County Communications Center, arrived on scene about 9:15 a.m., reported to the command post and offered assistance. About 10:15 a.m., OES arranged for two scenting dogs and their trainers to be flown from the San Francisco Bay area. The dogs and their trainers arrived about 3:55 p.m., and the trainers were briefed by the incident commander about the ongoing search and rescue efforts.

Meanwhile, about 2:00 p.m., SP began to set up blocks and tackle to facilitate removal of train debris with a crane. These efforts were halted by the incident commander about 3:00 p.m., before debris removal began, because the incident commander and the OES believed that such efforts might endanger rescue operations. The incident commander decided, and SP and Calnev representatives concurred, that nothing would be moved until the dogs had completed a search of the area.

The dogs alerted rescuers at various times when they sniffed the vicinity of the house at 2326 Duffy Street between 4:20 p.m. and 9:00 p.m. Shortly after 9:00 p.m., the rescue workers located a hand projecting through the debris at 2326 Duffy Street. The surrounding area was immediately stabilized. An opening was cleared by paramedics, who sent down oxygen and took vital signs of the trapped person. With the help of power tools, the resident was eventually freed from the debris about 10:34 p.m., about 15 hours after the derailment.

About 11:20 p.m., a rescuer was alerted by a dog in the vicinity of the third head-end locomotive unit. After removal of debris, the head-end brakeman was found dead in that unit about 3:03 a.m., May 13. The dogs worked until about midnight, examining all affected residences and portions of the train. By early morning on Saturday, May 13, the incident commander determined that all areas had been adequately searched, there were no further reports of missing persons, and, consequently, search and rescue efforts were terminated.

Shortly after noon on May 13, before wreckage removal operations began, SP bulldozers and hundreds of sandbags were used to build a dam at the lowest end of the accident site to help contain gasoline should the pipeline become compromised.

The San Bernardino Chapter of the American Red Cross initially learned of the train derailment on commercial radio about 8:43 a.m. At that time, representatives of the Red Cross responded to the scene where they met with the incident commander and were directed to prepare a shelter for 50 to 100 persons. The Red Cross Disaster Coordinator then contacted the Red Cross chapter office and requested additional personnel and logistical support. A temporary shelter was prepared at the local Job Corps building, a mobile canteen/kitchen was established at the accident site, and damage assessment teams were sent to the scene. The Executive Director for the San Bernardino Red Cross stated that they were equipped to handle the emergency and that they received logistical support from the Los Angeles and the Riverside Chapters in the form of a van, a canteen, and food supplies.

Wreckage Clearance/Pipeline Surveillance Operations Following the Train Derailment

May 12, 1989.—When Calnev’s manager of engineering received information regarding the train derailment, he radioed Calnev’s Colton terminal, about 6½ miles from the derailment site, and instructed personnel at that location to shut down the 14-inch pipeline immediately. At 8:30 a.m., pumping operations were stopped, leaving a residual pressure of 1,128 psig at Colton. The manager of engineering then notified Calnev’s manager of operations and the maintenance superintendent of the train derailment; all three individuals proceeded to the accident site to view the derailment and determine the potential impact to the pipeline. According to the manager of operations, when they arrived at the derailment site, it was obvious the pipeline could have been damaged because the pipeline was under a portion of the wreckage, "…most notably a locomotive that came to rest inverted directly over the pipeline" (figure 4). According to the manager of operations, their concern was that if the locomotive had remained intact, it could possibly have protruded into the ground 8 to 10 feet, and they were unsure at that time of the precise depth of the pipeline at that location. According to Calnev personnel, the derailment prevented Calnev from accessing the pipeline and performing any inspections of the pipeline in that location at that time. Calnev’s activities during the morning of May 12, according to the maintenance superintendent, were confined to remaining on site to make sure that no actions occurred on the part of the railroad or other agencies that could further endanger the pipeline. However, Calnev wanted to reduce further the pressure in the pipeline in the area of the derailment. According to the maintenance superintendent, "What we ideally were going to accomplish was to remove all of the product from the pipeline under the derailment area. As events proceeded, it was determined that that was unfeasible."

At 11:30 a.m., a foreman for Arizona Pipeline Company,^[13] permanently assigned to work on Calnev projects, arrived on site to assist Calnev personnel in reducing the pressure in the pipeline. The initial plan was to excavate the pipeline at a location 500 to 800 feet south of Highland Avenue (south of the derailment site), install a fitting for the purpose of tapping a hole into the pipeline, and withdraw product at that location. According to the Calnev maintenance superintendent, they were aware, by referring to company pipeline maps, that a check valve was installed in the pipeline immediately north (upstream) of the derailment site at pipeline milepost (MP) 6.9^[14] (figure 5). Calnev officials stated that they believed that removal of the product from the pipeline at the location south of Highland Avenue would cause the check valve to seat (close) thereby isolating the pipeline north of the check valve from the pipeline in the derailment area. Further removal of product from the pipeline would then reduce the pressure in the pipeline in the derailment area. After excavating at the location south of Highland Avenue, Calnev officials determined that the location was not suitable for tapping the pipe because the pipe was buried in the ground at a depth of 14 feet and was inside a steel casing. Calnev officials then moved their activities to the Colton terminal where a 2-inch fitting with a 1¼-inch opening was installed on the 14 inch pipeline, and they subsequently began withdrawing product from the pipeline at that location.

According to Calnev’s maintenance superintendent, after about 120 barrels of product were removed from the pipeline (and loaded into a vacuum truck), the pressure was reduced about 60 psig at the Colton pump station (MP 0.0) and at Cajon Pass (MP 25.7).^[15] Because the pipeline pressure had been reduced by an equal amount on both sides of the check valve at MP 6.9, Calnev personnel determined that they had not been successful in seating (closing) the check valve at that location and, consequently, had not been successful in isolating the pipeline in the area of the derailment. The equal reduction in pressure also indicated that the check valves at MP 14.9 and MP 19.2 had not seated.

Believing that they had been unable to withdraw product at a rate adequate to induce product backflow sufficient to fully seat the check valves, Calnev personnel installed a threaded fitting through the new opening and connected it with high pressure hoses in an attempt to withdraw product at a faster rate. According to Calnev personnel, a second vacuum truck load of product (120 barrels) was then withdrawn and comparable results were observed—an equal reduction in pressure on both sides of the check valve at MP 6.9. As a result, Calnev knew that the check valve at MP 6.9 was not closing. Calnev’s maintenance superintendent stated that he then recommended that additional pressure reduction could be achieved by closing the block valve at the Cajon Pass pump station. After the block valve was closed, a third vacuum truck load of product (120 barrels) was withdrawn from the pipeline and a 200-psig reduction in pressure was achieved. Once again, however, the pressure readings at the Cajon station and at the Colton station

Figure 5.—Elevation of Calnev pipeline.

indicated that the pressure had been reduced by equal amounts, which indicated to Calnev personnel that the check valves still had not seated. The 200-psig reduction also indicated that the remaining pressure on the line was due to the weight of the liquid and, as the maintenance superintendent stated, "that additional efforts would be only minimally successful in reducing the pressure at the Highland Avenue location [derailment site]," because backflow sufficient to seat a 14-inch check valve clapper could not be induced by withdrawing product through a 1¼-inch opening. As a result, Calnev suspended activities to reduce further the pressure on the pipeline, which at 10:00 a.m. on May 12, was 800 psig at Colton, or about 50 percent of the maximum operating pressure established by Calnev. According to Calnev’s manager of operations, Calnev did not at that time consider the possibility that the check valves were malfunctioning, but believed that the check valves did not close because of the inadequacy of the method used to induce backflow.

Meanwhile, SP’s division mechanical officer and other SP personnel had arrived on site and in consultation with Calnev and the incident commander began discussing plans for removal of the railroad equipment. According to the division mechanical officer, "the plan was to remove the cars and in no way affect the pipeline." The plan included cutting a breach (road) in the railroad levee through which the railroad equipment would be moved to the other side of the track. According to the San Bernardino Fire Department and Calnev, SP was advised that when the cars were to be removed, all cars were to be lifted and not dragged over the pipeline. Calnev’s manager of operations testified that he was aware of an accident in Montclair, California, in the latter part of 1988, during which wreckage removal operations possibly caused damage to a pipeline and that he wanted to avoid a repeat of such an incident. According to Calnev’s manager of operations, he did not discuss with the Fire Department or SP at that time what actions Calnev would take to inspect its pipeline after the cars were removed. Search and rescue operations continued until late in the evening on May 12, and efforts to begin removal of the wreckage were delayed until the following day.

May 13, 1989.—On the morning of May 13, SP removed 50 to 75 feet of track in preparation for making the breach (road) through the railroad levee that would be used for removing the railroad wreckage from the east side of the track to the west side. According to SP’s division mechanical officer, the site of the breach was determined by a break in the distribution of wrecked cars on the east side of the track (figure 4). Once the breach had been made, two 225-ton cranes and several bulldozers and front-end loaders came through the breach from the west side of the track, crossed over the pipeline, and were positioned at various points around the wreckage (figures 6 and 7). SP’s division mechanical officer testified that a lot of the trona that had spilled from the train was used to cover the ground and that with the the trona and the fill removed from the levee, there was about 6 to 7 feet of cover over the normal level of the ground in the area through which the equipment was moved. At the time the breach in the levee was made, the exact depth of the pipeline below natural grade had not been determined. During the morning of May 12, Calnev personnel used a line locator and yellow paint to mark the location of the pipeline throughout the derailment area. Later

Figure 6.—Equipment used during wreckage removal.

Figure 7.—Equipment used during wreckage removal.

that morning, with a backhoe and shovels, Calnev personnel dug two holes on either side of the locomotive engine that came to rest inverted over the pipeline and determined that the depth of the pipeline in that area was between 7 and 8 feet.

According to the testimony of Calnev’s maintenance superintendent and SP’s division mechanical officer, in removing the cars, the cranes would pick the cars up and swing them around in the breach in the levee. From that location, front-end loaders would then carry the cars to the west side of the track (figures 8 and 9). Testimony further indicated that equipment continuously operated through the haul road over the pipeline and that it was

Figure 8.—Equipment used to lift cars during wreckage removal.

Figure 9.—Equipment used to move cars to west side of track.

necessary on many occasions to re-mark the location of the pipeline with yellow paint. As Calnev’s maintenance superintendent testified, "…trona…was a very light, loosely compacted material…once you made a mark on it, it would take a very small amount of activity by heavy equipment to totally erase that mark."

SP’s removal of the wrecked cars, which were spread over a large area and stacked two and three cars high at some locations, continued throughout the day. A Calnev representative was on-site to monitor the operations and to keep SP personnel aware of the location of the pipeline. The incident commander kept fire engines and foam units on alert status with lines charged whenever a piece of wreckage was moved from a critical location over the pipeline. Calnev’s maintenance superintendent testified that it was his understanding that removal of the wreckage would proceed during daylight hours only. When SP continued their activities after dark, Calnev’s maintenance superintendent notified his supervisor who then returned to the site. After the situation was discussed with the incident commander and SP personnel, it was agreed that operations would be discontinued. Activities were halted about 11:00 p.m. that evening. The incident commander stated that he believed the cooperation exhibited by both Calnev and SP was exceptional.

May 14, 1989.—Removal of the rail cars resumed about 6:00 a.m. and continued throughout the day. Again, a Calnev representative was on site to monitor the operations and keep SP’s personnel aware of the location of the pipeline. According to SP’s division mechanical officer, the cars were removed "…in the manner in which the had been stacked…using two hooks with one crane. We picked them all straight up and then moved them out." He further testified that none of the cars were dropped in this process. He observed that debris including car components, axles, and pieces of rail remained in the area after the cars were removed; the visible debris was then also removed from the site. According to Calnev’s maintenance superintendent, it appeared that the debris had not penetrated the natural ground cover. SP’s division mechanical officer testified that no contact with the pipeline was observed during removal of the debris and "there was no rail sticking in the ground." Equipment operators working during the clearing of the train cars stated that many pieces of heavy construction and excavation equipment, including front-end loaders, cranes, and bulldozers worked simultaneously throughout the derailment area.

May 15 and 16, 1989.—When activity resumed on the morning of May 15, SP began making preparations to move the locomotives; all rail cars had been removed from the east side of the track. Calnev’s maintenance superintendent noted that the trona was scattered in varying depths throughout the area and over the pipeline to a point near, but not reaching, the engine (unit SP 7549) that lay inverted over the pipeline near the toe of the railroad embankment. To remove the locomotive units from the east side to the west side, SP personnel used two cranes to lift each unit and place it in the breach where one of the cranes, with the help of a front-end loader, carried the unit to the open field on the west side of the tracks. Each time a locomotive unit was moved, it was necessary for one of the cranes to cross through the haul road over the pipeline. Calnev personnel agreed that the crane could cross over the pipeline in this location. Calnev’s maintenance superintendent testified, "I did not see any activity which I believed damaged the pipeline. Any time you are using large pieces of excavating type equipment near a pipeline, you certainly have the potential for danger." According to SP’s division mechanical officer, who was in charge of the wreckage removal, he did not perform or know of any calculations that were performed to determine the stress imposed on the pipeline due to the weight of the cranes and the cars that were carried across it.

When the locomotive that came to rest inverted over the pipeline was removed by SP, Calnev personnel observed that the entire top of the locomotive had been sheared off and that it had been resting at grade level. There was nothing visible protruding into the ground. Calnev, however, decided to excavate the portion of the pipeline that had been under the locomotive. Using a backhoe equipped with a 24-inch bucket, Calnev personnel excavated an area approximately 80 feet in length parallel to and about 2 feet east of the pipe to a depth about 4 inches lower than the depth of the pipe in the area. Pipe depth was reported to have been about 8 feet at the southern end of the excavated area and 6½ to 7 feet at the northern end. According to Calnev personnel, the soil surrounding the pipe was removed by hand so that the pipe was exposed from the 6 o’clock position to the 2 o’clock position facing south (see figure 4, excavation # 1). Calnev’s manager of operations testified that he personally entered the excavation, inspected the pipe, and found no damage to the coating or to the pipe.

Calnev officials then decided to excavate in an area north of the breach where, according to Calnev’s manager of operations, "…bulldozers had been repeatedly going off the end of the haul road" (figure 4, excavation # 2). According to the Arizona Pipe Line Company foreman, who performed the excavation, about 1 foot of pipe length was exposed from the 1 o’clock to 3 o’clock position looking north. When asked if any damage to the coating or pipe was noted, the foreman replied, "Couldn’t really tell by a visual look, and we didn’t bother exposing anymore due to our objective was to determine depth and alignment of the pipeline at that time." The depth of the pipe at this location was determined to be about 7 feet. With respect to the depth of the pipe, Calnev’s manager of operations testified, "…it was sufficient to where I was no longer concerned about any damage from the bulldozer activity."

By late afternoon on May 15, the wreckage had been removed and SP began to demolish the houses that had been damaged during the derailment. SP planned to close the breach that evening, relay their tracks, and begin removing the trona on the following day, May 16. According to Calnev officials, it was at this point that they began to formulate the next step of their inspection plan. Calnev understood that if SP began removing the trona on Tuesday, inspection of the pipeline would be delayed until the trona removal was completed. According to Calnev’s manager of operations, "At that point, we were still unsure of the integrity of the pipeline. It was still in a stable situation. It had not lost any pressure nor were there any signs of leakage. But yet we could not verify the integrity of the pipeline before then." Calnev’s plan was to move in additional equipment, remove all the trona over the pipeline down to native soil, and excavate and inspect the pipeline at any location where debris was found and appeared to have penetrated the native soil. According to Calnev officials, by removing the trona from over the pipeline, SP personnel would not have to work directly over the pipeline when they began hauling away the trona on the following day. According to Calnev’s manager of operations, this plan was discussed with SP officials and the incident commander, and no recommendations or modifications to the plan were suggested.

Using a John Deere 690B excavator and working from south to north, Calnev began making a path about 8 feet wide through the trona beginning at a point near where the locomotive came to rest inverted over the pipeline (figure 4). According to Calnev’s maintenance superintendent, the excavator was followed by a front-end loader to complete the removal of the trona. He further testified that a few inches of natural soil was removed and that as much as 12 to 16 inches may have been removed at any one point, but that he still believed that he had plenty of cover over the pipeline.

In making the 8-foot-wide path, Calnev piled the trona that was removed from over the pipeline to the east of the pipeline at a distance, estimated by Calnev’s manager of operations, to have been 2 to 4 feet. He testified, however, that "we found that the trench [path] did not place the pipeline right in the middle. There was an area where the pipeline kind of hugged the side of the trench [path], so it [pile of trona] could have been as close as 2 feet in that area."

Calnev’s maintenance superintendent, who supervised the trona removal activity from about 8:00 p.m., on May 15, to about 4:00 a.m., on May 16, testified that several pieces of debris, including portions of truck assemblies [from a train car] and two pieces of rail—one about 3 feet in length and one about 10 feet in length—were found during removal of the trona. He further testified that while he was supervising the removal of the trona, two excavations of the pipeline were performed where debris had been found at natural grade level. He stated that he could not be specific about the locations but estimated that the first excavation was near the north edge of lot 77 and that the second excavation was between lot 77 and lot 76 (figure 4, excavations # 3 and 4). For both excavations, the depth and the alignment of the pipe were determined by digging with hand shovels. A Case 580C backhoe was then used to excavate on the east side (Duffy street side) of the pipeline. According to the maintenance superintendent, no damage to the coating or the pipe was observed.

SP personnel had positioned lights on the railroad levee. According to Calnev’s maintenance superintendent, even though the lighting cast shadows in the excavated area from west to east, lighting was not an issue in determining whether the pipeline had been damaged or in evaluating the depth of cover over the line. He stated, "I was comfortable with the level of lighting, and I spent a considerable amount of time in the trench closely observing the excavation." He also testified that it would have been possible to detect the difference between hitting debris with the backhoe and hitting the pipeline with the backhoe. "…it was never a concern of mine that we were going to hit the pipeline with the backhoe because we were monitoring the depth of cover over the pipeline. We were not excavating in an area such that we would be getting close enough to the pipeline to hit it."

In addition to the two excavations, the pipeline was potholed^[16] at several other locations. At one location where the pipeline was potholed, a truck assembly [rail car] was found to have penetrated the natural soil. Calnev’s maintenance superintendent marked this location and later advised Calnev’s manager of operations of the need to perform a more thorough inspection of the pipeline at that location. By 4:00 a.m., on May 16, the path through the trona had extended north 300 to 400 feet to a point where the breach in the levee had been made.

The deputy fire chief testified that when he terminated his role as incident commander around 10:00 p.m. on May 15, Calnev’s manager of operations assured him that the pipeline was safe to operate.

Calnev’s manager of operations, who relieved the maintenance superintendent about 4:00 a.m. on May 16, supervised the remainder of the trona removal from over the pipeline. A foreman for Arizona Pipe Line Company arrived on site about 6:00 a.m. and relieved the backhoe operator who had worked through the night. According to Calnev’s manager of operations, two additional excavations of the pipeline were performed; he estimated the first excavation to be near the middle of lot 76 (figure 4, excavation #5), where the maintenance superintendent earlier had found a truck assembly, and the second location to be near the northern edge of lot 75 (figure 4, excavation #6). At both locations, the excavation was performed on the west side of the pipeline, a 20- to 25-foot section of the pipe was exposed from the 6 o’clock position to the 2 o’clock position facing north, and no damage to either the coating or the pipe was observed by Calnev personnel. The depth of pipe was determined to have been about 4 feet at the first location and 5 feet at the second location.

According to the testimony of Calnev officials and the backhoe operator, all the excavations were immediately backfilled after the coating and pipe were inspected for damage. Further testimony indicated that about 6 inches of debris-free native soil would be used to fill the remainder of the excavations, and that compaction of the soil was accomplished by "wheel-rolling" rather than by use of the backhoe bucket.

Beginning about 10:30 a.m. on May 16, Calnev began performing soft dig excavations^[17] of the pipeline about every 50 feet throughout the derailment area. At each location, an 8-foot-tall stake marked at 1-foot intervals was placed on top of the pipe, the top of the stake was surveyed to determine its elevation, and the hole was backfilled. Calnev personnel testified that as a result of these soft dig excavations, the pipe was exposed from the 10 o’clock position to the 2 o’clock position at each soft dig excavation and that before the holes were backfilled, the pipe was inspected for damage; no damage was observed at any of these locations. According to Calnev, the purpose of the stakes was to provide information to SP regarding the location and depth of the pipeline when SP began removing the trona from the derailment site. SP was advised by Calnev to preserve the stakes until all grading of the area was completed. Calnev’s manager of operations observed, based on the placement of the stakes, that the pipeline depth below natural ground varied from 4 to 8 feet through the derailment area.

Calnev’s manager of operations testified, "On Tuesday, the 16th, we had by then accomplished full trenching [8-foot-wide path] over the top of the pipeline in the affected area. We had removed or had caused to remove any debris that we had found. We had investigated every area that debris had penetrated the native soil. …Based on that assessment…my opinion was that the pipe had not been damaged by the train derailment." Clearance was given at 11:28 a.m. by Calnev for the restart of the pipeline; operations were resumed about noon on Tuesday, May 16. The pressure was initially increased to about 1,200 psig, at which point, according to Calnev’s manager of operations, the dispatcher on duty watched for any signs of loss of pressure in the system. The pressure held constant for about 15 minutes after which the pipeline was brought up to normal operating pressure (about 1,600 psig) and regular operations were resumed.

The Safety Board received conflicting testimony regarding a request to expose completely the pipeline prior to resuming operations. The incident commander (San Bernardino deputy fire chief) testified he requested that Calnev fully expose the pipeline in the derailment area. According to Calnev’s manager of operations, such a request was not made by either the San Bernardino fire department or the SP. He did state that several options had been considered, including the use of an internal electromagnetic inspection instrument for detecting defects in the pipe wall and a hydrostatic test of the pipeline. He stated further that it would not have been practical to run the inspection instrument through the line because "…the line would have had to have been brought up to full operating pressure and operated in that state for about 5 days to push [the instrument] through to the other end." He elaborated that because of the mountains between Colton and Las Vegas [the end of the line], it would be necessary to operate at full pressure just to get the instrument over the mountains. Calnev’s manager of operations also stated that, "[A] hydrostatic test would have been performed had there been some doubt as to the integrity of the pipeline. We found no reason to doubt the integrity of the pipeline upon completion of our inspection and did not perform a hydrostatic test."

SP contracted with the International Technology Corporation (IT) to have the trona removed from the derailment site; removal of the trona began during the afternoon of May 16. According to the project manager for IT, cleanup of the trona began in the area closest to Duffy Street and then continued through the derailment area from south to north. Equipment operators testified that to remove the trona that had been piled east of the pipeline as a result of the 8-foot-wide path that had been made through the trona, the operator of a front-end loader would reach over the pile of trona with the bucket of the loader and drag the material back toward Duffy Street where the trona could then be loaded into trucks. According to the IT manager, the front-end loader worked perpendicular to the pipeline during this operation.

At 4:00 p.m. on May 16, SP opened its line to resume train movements through the area.

May 17, 18, and 19, 1989.—Removal of the trona continued throughout the day on May 17 and 18. Because trona contrasts with the color of the native soil, operators of the equipment were told by IT to visually inspect the area to assure that they had removed all of the trona and about the top 2 inches of native soil. On May 18, a track-mounted (crawler type) excavator was brought to the site to begin removing the trona from the railroad embankment. The excavator was positioned east of the pipeline with the tracks parallel to the pipeline. A smooth steel grading blade was welded to the teeth on the bucket of the excavator. The blade enabled the operator to drag trona that was covering the railroad embankment without removing excessive amounts of material and to leave behind a smoothly graded surface. Testimony by equipment operators in the area at this time indicated that the operator of the excavator would drag the trona down the side of the railroad embankment and across the pipeline to the east side where front-end loaders would pick up the trona and load the trucks. However, according to IT’s project manager, the operator of the excavator would drag the trona down the embankment and build a stockpile of trona on the west side of the pipeline. At that point, a front-end loader would come in, keeping the tires on the east side of the pipeline, scoop up the material, and then back up to a point where the material could be loaded into trucks. Testimony by equipment operators further indicated that the smooth-edged blade welded to the teeth on the bucket of the excavator broke off several times and that the equipment continued to be operated without the smooth-edged blade. According to IT’s project manager, the excavator made two "passes" on the embankment, one pass from south to north and one from north to south.

By early afternoon on May 19, 1989, all the trona had been removed and the fencing of the area that began during the morning was completed. The last piece of equipment used for the cleaning operations, a motor grader, was brought to the site to smooth out the surface and remove tire tracks. After this operation was completed at 6:00 p.m., locks were placed on the two 20-foot-wide gates that were installed with the fence, and the area was secured. According to SP’s contractor, no equipment was used in the area after May 19, 1989.

IT’s project manager testified that when he left the site on May 19, he believed that there were 2 to 3 feet of ground cover over the pipeline. When asked, "Could it have been your work that removed that cover from the 4 to 8-foot level down to the 2 to 3-foot level?" He replied, "Yes."

According to Calnev, a Calnev representative was on site through May 19, during the removal of the trona, to observe the operations, to point out potentially dangerous situations to the railroad and its contractor, and to make certain that the stakes that had earlier been located over the pipeline remained in place. No concern was voiced by Calnev during the removal process.

Events Preceding the Pipeline Rupture

Calnev’s dispatch center at the Colton Pump Station is equipped with a monitoring system that scans and records, among other system parameters, pipeline pressures. When normal operations resumed on May 16, the pipeline pressure had increased to 1,667 psig. Between May 16 and May 23, the pipeline was operated at pressures ranging between 1,690 and 1,060 psig (normal operating ranges established by Calnev) and was subjected to various pressure changes during this time. Operations during the next couple of days showed only smooth pressure transitions until about 8:05 a.m.^[18] on May 25, 1989.

Pipeline Rupture

Pipeline Operations on May 25, 1989.—During the early hours of May 25, 1989, the three 1,000 horsepower (hp) mainline pumps at the Colton Terminal were operating at maximum output (2,300 to 2,400 barrels per hour), and the pressure on the pipeline was relatively constant at 1,620 psig. About 4:03 a.m., with the completion of a product delivery at Daggett (see figure 1), a gradual increase in pressure to 1,680 psig occurred over an interval of about 17 minutes at which time the pressure decreased within 5 minutes to 1,669 psig. The pressure then remained relatively constant until 8:05 a.m.

At 8:05:25, based on a readout of the information recorded by the monitoring system, a low suction pressure (15.188 psig) alarm^[19] and a low discharge pressure (257.644 psig) alarm were received in the dispatch center at Colton Pump Station on Calnev’s computer system. At 8:05:38, the three 1,000-hp mainline pumps were shut down by the computer system. At 8:05:39, the dispatcher acknowledged^[20] the alarms. According to testimony of the dispatcher on duty at the time, when changes in operating conditions occur: (1) an audible alarm will be sounded, (2) the word "alarm" will appear and flash at the top of the dispatcher’s computer terminal screen, and (3) information regarding the specific condition (in this case, "low suction pressure" and "low discharge pressure") will be highlighted in a particular color and continue to flash until acknowledged by the dispatcher. Testimony further indicated that if more than one condition occurs on the same page [screen], the word "alarm" and the audible alarm are terminated by one stroke on the computer keyboard.

The dispatcher testified that he noticed on his terminal screen flashing lights indicating that the pumps were shutting down and that he had a "low suction pressure" color alarm (blue). He did not notice the "low discharge pressure" color alarm (blue) on the same page. The dispatcher stated that he believed the pumps were shut down as a result of a low liquid level in the storage tank from which he was pumping. He was aware that a similar situation had been experienced by the dispatcher whom he relieved, and the pumps were eventually restarted. According to the dispatcher, the normal procedure for the condition of a low liquid level in a storage tank is to restart the pumps after the suction pressure again returns to normal. According to the dispatcher, normal suction pressure is between 26 and 50 psig. The suction pressure rose to 37.1429 psig, and at 8:06:02, the dispatcher commanded the restart of the 100-hp booster pump. At 8:06:11, the command was acknowledged by the computer. At 8:06:22, the computer reported the status of the booster pump^[21] as "off."

At 8:06:53, the dispatcher again commanded the computer to start the booster pump, and at 8:06:57, the command was acknowledged. Operating parameters were automatically checked and found satisfactory, and the system attempted to restart mainline pumps Nos. 2 and 3. At 8:07:09, the computer acknowledged the command. At 8:07:19, another low suction pressure (17.2932 psig) alarm was given to the dispatcher who knowledged the alarm, and at 8:07:22, mainline pump No. 2 registered status "off," as did mainline pump No. 3 at 8:07:23. Also, at 8:07:23, the suction pressure was 46.1654 psig and at 8:07:55, the booster pump reported status "off."

At 8:08:10, the dispatcher acknowledged the shutdown alarms and again commanded the start of the booster pump. At 8:08:18, the booster pump acknowledged the command and at 8:08:19, pump No. 3 acknowledged the command. At 8:08:20, a low suction pressure (20.9023 psig) alarm was provided to the dispatcher. Pump No. 3 reported status "off" at 8:08:32, at which time suction pressure was recorded as 90.9774. At 8:09:15, the booster pump reported status "off." At 8:09:18, the shutdown was acknowledged by the dispatcher. The dispatcher stated that because he was not successful in restarting the pumps, he left his station to request assistance from another dispatcher who was on duty as a supervisor at the time and located down a hallway from the dispatch center. The supervisor acknowledged the request.

While returning to his dispatch area, the dispatcher encountered the senior systems specialist and asked him if he knew of any reason why the pumps would not come back on. The dispatcher stated that the systems specialist advised him to "pinch down" on the station control valve to bring the pumps on slowly. The dispatcher stated that as he was doing this, they received a phone call from the San Bernardino County Communication Center asking if Calnev’s pipeline was involved in a fire. The systems specialist then observed through a station window a cloud of smoke in the direction of the pipeline route through San Bernardino, advised the caller that it likely was Calnev’s pipeline, and then instructed the dispatcher to leave the pumps down.

After notifying Calnev locations currently taking delivery of products at Las Vegas, Nevada, that the pipeline was being shut down, the dispatcher began remotely closing valves to isolate the pumps and the storage tanks from the pipeline. In addition to closing the valves at the terminal, he shut down the Baker booster pump station at MP 146.2. After the pressure sensor indicated zero psig pressure at the summit of Cajon Pass, the dispatcher remotely closed the valve at California aqueduct (MP 35.4) which is located on the north side of Cajon Pass. He also stated that notification was made to personnel who had to close other valves manually. The first downstream valve that had to be closed manually was located at MP 25.7; the maintenance supervisor reported that this valve was closed at 9:00 a.m.

Witnesses’ Observations.—A resident at 2395 W. Adams Street stated that she was in her backyard between 7:45 a.m. and 8:00 a.m. and noticed a "white colored rain" falling on the house behind hers on Duffy Street. She further stated that after she went back inside her house, she heard an explosion and "then her windows blew in" and the entire house was on fire. Another resident at 2446 San Benito Street stated that he was outside around 8:00 a.m. on May 25, heard a train go by, and about 5 to 10 minutes later heard a "rumble." He stated that he then looked up and saw a "cloud of flame about four houses wide come over the houses…the flame was about 10 feet higher than the rooftops" (figure 10). Several witnesses stated that they saw a white vapor and then heard a loud explosion; this was followed by black smoke and intense heat and flames. A resident at 2385 Mesa Street recalled that a friend, who had arrived at her residence to transport her children to school, "pointed to a spray vapor shooting up into the sky," that was coming from the direction of where the train had derailed. A motorist, who was filling his automobile with gasoline near Macy Street and Highland Avenue, stated that he heard a "rumble," then saw what appeared to be a "geyser" of liquid shooting up in the air near the site of the train derailment. He stated further that within a few moments "it exploded." In addition to the resident on San Benito Street, several residents recalled hearing a train pass by 5 to 10 minutes before the explosion; residents also recalled smelling gas before the explosion. Two residents, one at 2327 Duffy Street and one at 2315 Duffy Street, were fatally burned as a result of the explosion and fire.

Emergency Response to Pipeline Rupture

On May 25, 1989, at about 8:00 a.m., a firefighter leaving his assigned fire station on Highland Avenue noticed a large column of black smoke in line with Highland Avenue, about 2 miles from his location. He returned to the fire station and notified the battalion chief.

Figure 10.—Fuel burning after the pipeline rupture.

The battalion chief, in turn, notified his dispatch office about 8:08 a.m. and requested fire department personnel and equipment to respond to Highland and Duffy Streets. En route to the site, the battalion chief observed flames and black smoke rising straight up in the air with no apparent wind. He arrived on-scene about 8:13 a.m. Mutual aid agreements were activated when the dispatch center was notified of the accident. As emergency response units and fire department personnel and equipment from adjacent jurisdictions arrived on scene, the battalion chief positioned them around the involved area. He had surveyed the accident area and determined that seven houses were fully engulfed in fire and that two houses were partially on fire. Being concerned with the downed power lines and the possibility of ruptured residential gas lines, the battalion chief requested the utility companies to shut down their respective lines. He also requested the water department to assist in building dikes to prevent the product from flowing into surrounding areas. The battalion chief ordered an evacuation of residents in the area; police personnel eventually evacuated about 170 persons in a four-block area. According to the deputy fire chief, because of fuel remaining on the ground, some residents were unable to return permanently to the area until August 6, 1989.

At 8:30 a.m., the deputy fire chief, who had been the incident commander during the response to the train derailment, arrived on scene and assumed the role of incident commander for this accident. By the time he arrived, fire-fighting operations and treatment and transportation of the injured to local hospitals had begun. At 10:05 a.m., a command post was set up at 2359 Mesa Street. According to testimony of the deputy fire chief, the mutual aid emergency response plan was implemented as planned. Although the deputy fire chief’s role as incident commander ended on May 28, fire department personnel remained on scene as a safety measure until May 31, 1989.

Pipeline Surveillance Operations

After Calnev’s maintenance superintendent observed the fire from his office window shortly after 8:00 a.m., he immediately notified the manager of operations who, along with other company personnel, proceeded to the accident site. Upon arrival at the accident site, the manager of operations introduced himself to the incident commander and was directed by the incident commander to fly with a police officer in a helicopter to observe the fire. Calnev’s manager of operations stated that while in the air, he observed a large stream of flaming liquid exiting the ground eastward at an angle of about 60 degrees from the horizontal. He stated that he observed substantial fire damage in the direction of the burning stream of liquid, a small pool of liquid burning around the rupture, and a small grass fire burning south of Highland Avenue. The manager of operations stated that he then advised the incident commander to allow the fire to burn itself out. According to the incident commander, the fire burned out by 3:30 p.m. on May 25.

According to Calnev’s manager of operations, when the fire was out, the rupture site was inspected and the damaged pipe examined (the damaged is described in the section "Damage," "Damage to the Pipeline"). At least four pieces of railroad debris—a brake arm, an approximately 8-inch section of I-beam from a locomotive, a piece of metal cowling from a locomotive, and a short section of rail—were found near the rupture. The brake arm and the rail section were about 2 to 3 feet in length. The brake arm was found 8 inches above the pipeline and the other parts were within 2 feet of the pipeline. Testimony by Calnev’s manager of operations and by equipment operators who had worked at the site following the derailment indicates that the depth of cover they observed over the pipeline at its point of rupture was from 2 to 2½ feet, whereas the depth of cover they had observed after completing work, following the train derailment, was from 4½ to 6 feet. Calnev’s manager of operations testified that the location of the rupture was very near if not at the exact location where the excavation #5 had been performed across from the middle of lot 76 (figure 4).

According to Calnev’s manager of operations, Calnev’s plan to repair the pipeline after the rupture and place it back in service evolved over many days "…during which many discussions were held with many interested parties as to how best to return that pipeline to service [and] make the repairs necessary." Calnev’s maintenance superintendent testified that when the pipeline rupture occurred, he notified the National Response Center, the California Office of Emergency Services, the California State Fire Marshal’s Office, and the Underground Service Alert System. Representatives from these agencies, as well as an engineer from the U.S. Office of Pipeline Safety (OPS), responded to the accident site.

On May 26, 1989, OPS issued a Hazardous Facility Order, CPF No. 5987-H to Calnev (appendix E). This Order included preliminary findings, among others, that the pipeline within the area of the derailment had not been completely exposed and visually examined for damage, that the structural integrity of the portion of the pipeline potentially affected by the derailment had not been ascertained by Calnev, and that Calnev had not determined if there had been damage to the pipe coating as a result of the cleanup operations. OPS required Calnev to expose the pipeline around its circumference for the length of pipe between a point 50 feet north of the casing beneath Highland Avenue and the south end of the levee adjacent to the check valve at MP 6.9, to conduct a thorough visual inspection of the exposed pipe to locate any damage to the pipe or to its coating and make appropriate repairs, and in accordance with applicable requirements of 49 CFR Part 195, to hydrostatically test to 1.25 times its maximum operating pressure the pipe located between a point 100 yards south of the check valve on the downstream side of the derailment impact area and a point 200 yards upstream of the road crossing at Highland Avenue.

On May 30, 1989, based on its preliminary findings, OPS found that "if placed into service under the same circumstances as existed after the rupture, that portion of Respondent’s [Calnev’s] pipeline subject to the required corrective actions prescribed [on May 26, 1989] would be hazardous to life and property." Consequently, as a result of conversations with Calnev, the OPS order was amended to require excavation of the pipeline located between a point 10 feet north (downstream) of the casing beneath Highland Avenue and the south (upstream) rise of the Muscoy Levee, that the excavated pipe be visually inspected to determine any damage to the pipe or its coating, that the pipe be replaced with new pipe, that a block valve be installed between the check valve and the Muscoy Levee, that the new pipe be tested as previously required of the existing pipe, and that all activities be performed in accordance with applicable requirements of 49 CFR Part 195. The revised Order also stipulated that OPS would review and approve Calnev’s hydrostatic testing and inspection program, that OPS would monitor the test, and that the pipeline could not be returned to service until OPS had determined that all required actions had been successfully completed.

By letter of June 6, 1989, Calnev requested relief from the requirements of the Order because it discovered that a bend in the pipe made it impractical to tie into the new pipe 10 feet north of the Highland Avenue casing, the location required by the amended Order. As there was no apparent damage to the pipe at that location and because the line would be hydrostatically tested before returning it to service, on June 6, 1989, the OPS again amended the Order to allow the tie-in to be made at a location determined acceptable by Calnev and concurred with by a representative of the OPS so long as the tie-in was made between the point 10 feet north of the Highland Avenue casing and a point about 35 to 40 feet north of the casing.

About 600 feet of the pipeline through the area of the previous derailment was removed and replaced. The pipeline was refilled with product on June 9, 1989. More than 9,400 barrels of product were required to refill the pipeline. A mile of pipeline of the size installed will hold 917.69 barrels of product, based on information provided by Calnev.

Injuries

	Train Derailment		Pipeline Rupture
Injuries	Extra 7551 East	Residents	Residents	Firefighters	Others[22]	Total
Fatal	2	2	2	0	0	6
Serious	1	1	3	0	1	6
Minor	2	0	16	1	4	23
None	2	—	—	—	—	2
Total	7	3	21	1	5	37

Damages

Train Derailment.—Five locomotive units and the entire consist of 69 hoppers cars were completely destroyed as a result of the derailment; the rear-end locomotive was extensively damaged. About 680 feet of track were destroyed by the derailing locomotive units and cars.

Following the derailment, a building inspector from the City of San Bernardino Department of Building and Safety inspected the houses that incurred damages as a result of the derailment. The inspector’s assessment of damages is listed in Appendix F. The inspector recommended that dwellings at 2314 Duffy Street through 2382 Duffy Street be demolished immediately (figure 11).

Figure 11.—Damages to residences.

Pipeline Rupture.—Eleven houses and 21 motor vehicles were destroyed by fire from the pipeline rupture and fire (figures 11 and 12). Four houses received moderate fire and smoke damage, and three houses received smoke damage only. Appendix F lists the residences and the damages incurred.

The costs incurred from the train derailment and the pipeline rupture, as reported by SP, follow:

Equipment
⁠69 Cars	$ 1,550,407.00
⁠5 Locomotives	7,506,000.00
⁠1 Locomotive repair	85,001.00
Track	14,922.00
Wreckage Clearing	1,968,867.00
Lading	242,830.00
Houses
⁠Derailment (9)	592,831.00
⁠Rupture (7)	453,433.00
Total	$ 12,414,291.00

* The dollar figure is based upon comparable locomotives available today for replacement.

Calnev reported the following costs as a result of the pipeline rupture:

Pipeline Repair	$ 500,000.00
Commodity	300,000.00
Environmental	1,060,000.00
⁠Total	$ 1,860,000.00

Total reported costs from the train derailment and the pipeline rupture were: $14,274,291.00.

Damage to the Pipeline.—The 14-inch-diameter pipeline ruptured at about MP 6.9. A 25-foot, 1⅞-inch-section (301⅞ inches) of the pipe that included the rupture area was cut from the pipeline to make a temporary repair. The 25-foot section was removed about 5:00 p.m. on May 26, 1989, and was replaced with a section of sound pipe.

After the 25-foot section of pipe containing the rupture was removed, it was torch cut into 5 smaller sections. Beginning at the south end, the first section was 44⅞ inches long and contained no areas of damage. The next 44½-inch section contained the rupture. These last two sections of pipe were taken to the Safety Board’s laboratory in Washington, D.C., for testing (figures 13 and 14). (Additional information is provided under "Tests and Research.") The next section was 6 inches long and contained no damage. The last section was 165½ inches long and contained damage to the coating along the side of the pipe at the 3 o’clock position (looking north).

Figure 12.—Damages to residences following pipeline rupture.

Figure 13.—Section of pipe removed from rupture area.

Figure 14.—Section of pipe containing the point of rupture.

The rupture was about 29 inches long and, with respect to the circumference, was located about 5 inches to the east of the top of the pipe as installed (about the 1:30 o’clock position looking north). The electric resistance welded seam was located about at the top of the pipe. Examination of the area indicated that there was plastic deformation (bulging of the pipe) associated with the rupture and that the rupture produced a "fish mouth" type opening of about 4.2 inches at its widest point (see figure 14). There was no apparent visual evidence of pipe material or manufacturing" deficiencies.

Plastic deformation (denting) was present in the area of the rupture. The primary "dent" extended approximately 27 inches longitudinally along the top portion of the pipe; the dent angled slightly from the longitudinal axis of the pipe. The primary dent began at a point 20 inches northwest of the rupture point and extended to a point 7 inches southwest of the rupture point. The primary dent was about 1⅜ inches wide at its widest point and the deepest depth of the dent was abou 0.07 inches.

The primary dent produced a protrusion (bulge) on the inside surface of the pipe and localized wall thinning. The minimum wall thickness, as measured in this area at the accident site, was 0.249 inches and was located about 4 inches from the point of rupture. Additional wall thinning was near the point of rupture ("Tests and Research," "Metallurgical Testing").

Nearly parallel to and below the primary dent was a mark/scratch on the pipe that extended from about the same downstream location as the primary dent to about the point of rupture.

A second pair of marks on the pipe was located upstream (south) about 36 inches south of the point of rupture. The pipe had been damaged (gouged) in an area about 5 inches below (east) the top of the pipe. The longer mark was about 36 inches long and located closer to the top of the pipe; a 2½-inch-wide section of the coating had been removed. revealing a 1½-inch-wide mark on the metal. The second mark began slightly north of the first; the maximum width of damage to the coating was about 2 inches and the length of damage was about one half that of the upper mark.

About 120 inches north of the point of rupture, some damage to the coating on the east side of the pipe was observed. Coating in widths varying from 4 to 7 inches had been removed from the pipe at the 3 o’clock position (looking north). No damage was apparent o the pipe metal.

At a location 188 feet north of the end of the Highland Road casing, two reas of damage to the pipe were found at the time the pipe was being removed for replacement. The section of pipe containing these two areas of damage were sent to the Southwest Research Institute for metallurgical examination. ("Tests and Research," "Metallurgical Testing").

Track and Signal Information

Track.—The train derailment occurred on the single mainline track in San Bernardino, California, near railroad MP 486.8. Approaching the derailment site from the west, the track grade descended between 2.0 and 2.2 percent for 22 miles before it transitioned to a O.O-percent grade at the derailment site. In the 22 miles of descending grade, there were 56 curves which varied in degree of curvature from a maximum of 6 degrees to a minimum of 30 minutes.

The track was constructed of 119-pound continuous welded rail (CWR) on tangent track and 136-pound CWR on most of the curves. The 4-degree curve at the derailment site was laid with new 136-pound CWR in 1986. The rails rested on double shouldered tie plates and 9-foot hardwood. crossties and were restrained with two rail-holding spikes on the gage side of the rail, one rail-holding spike on the field side of the rail, and one plate-holding spike on each side of the rail. The rail was box-anchored every other tie. The ties were laid in a ballast of crushed rock.

The 4-degree right-hand curve (based on the direction of movement of Extra 7551 East) at the derailment site was constructed on a fill (levee) with a maximum height of about 21 feet. The curve was 2,474 feet in lenqth, including a 376-foot spiral on each end, and had a 1-inch superelevation.

According to SP Timetable No. 2, the authorized maximum timetable speed for the curve was 30 mph. The Federal Railroad Administration allows a maximum operating speed of 38 mph for a 4-degree curve with a 1-inch superelevation.

About 680 feet of track were destroyed during the derailment. Because of the extensive track damage, there were no distinguishable marks to indicate the point of derailment.

Signals.—Trains on the single mainline track are governed by a traffic control system using colored lights on wayside signals. An inspection of the signal equipment in the area of the derailment was conducted on May 13, 1989. The inspection revealed no problems with the signal system.

Train Information

At the time of the accident, Extra 7551 East consisted of, from front to rear, 4 road locomotive units (SP 8278, SP 7551, SP 7549, and SP 9340), 69 open-top hopper cars loaded with trona, and 2 helper locomotive units (SP 8317 and SP 7443).

Locomotive Units.—All of the locomotive units were manufactured by the Electro Motive Division (EMD) of General Motors Corporation. These units were six-axle, SD models with 26L automatic brake valves and extended range dynamic brakes.^[23]

Train-brakes were controlled by the road engineer in the lead unit, SP 8278. Dynamic and independent brakes were controlled separately by each engineer in their respective units, helper or road locomotive.

Based on statements by the head-end and helper engineers, the dynamic brakes of Units SP 8278 and SP 7443 were known to be functioning. Unit SP 7551 was dead-in-consist with no dynamic brakes or power. The dynamic brakes of unit SP 8317 were tagged and out of service, but the unit pulled in the power mode and had pneumatic brakes. The head-end engineer stated that he believed "the third unit (SP 7549) had fairly good, I think they were good dynamics." The event recorder printout for SP 7549 did not show any amperage in the dynamic mode after the train departed Oban where the helper units were added. The fourth unit, SP 9340, was reported by the head-end engineer to load in and out of dynamics intermittently.

The automatic and independent brake valves from units SP 8278 and SP 7443 were bench tested on May 15, at the SP diesel shop in Los Angeles in accordance with the requirements of the manufacturers and the Association of American Railroads. All valves performed within design specifications.

The controlling locomotive units at the head end and rear end of the train, SP 8278 and SP 7443 respectively, were equipped with multi-channel radios that broadcast on 30 watts of power at 72 volts. The road channel was 161.550 MHz. Both radios Were bench tested on May 14 and 15, at the SP radio facility at Colton Yard. Both radios functioned according to design and Federal specifications (49 CFR Part 90). On May 12, an on-scene functional test of the radio from SP 7443 transmitting to the Colton roundhouse was performed; communication was loud and clear.

The first three head-end locomotive units of Extra 7551 were equipped with Pulse 8 event recorders; the fourth head-end unit and the helper units were not equipped with any event or speed recorder. None of the units were required to be equipped. According to SP’s general road foreman, all new locomotives being purchased are equipped with event recorders, and event recorders are being installed on existing locomotives during a major overhaul. The helper units had not recently been through a major overhaul maintenance program. The Pulse 8 event recorder cartridges record speed, time, distance, direction, amperage, braking, throttle position, and independent brake application. All three event recorder cartridges were recovered and taken by Safety Board personnel to its headquarters in Washington, D.C., for restoration (the cartridge from unit SP 8278 was heavily damaged) and printout . (See "Tests and Research," "Event Recorders.")

Hopper Cars.—Of the 69 open-top hopper cars in the consist of Extra 7551 East, 38 cars were owned by the SP. These cars were 48 feet 9 inches in length, had a light weight of 60,300 lbs, a maximum lading capacity of 202,700 lbs for a maximum weight per car of 263,000 lbs. The remaining 31 cars were owned by the Denver & Rio Grande Western Railroad (DRGW). These cars were 51 feet 8 inches in length, had a light weiaht of 63,500 lbs, a maximum lading capacity of 199,500 lbs for a maximu weight pr car of 263,000 lbs. The total light weight of the 69 cars was 2,130 tons.

Each of the SP cars was equipped with an "empty load" (EL) device. When the car is empty, this device reduces the brake cylinder pressure to prevent the wheels from sliding. According to timetable instructions in effect at the time of the derailment, loaded cars with empty load devices were to be considered the equivalent of one and one-half cars in determining tons per operative brakes (see Southern Pacific’s Method of Operation). The chief mechanical officer for SP testified that the SP cars with empty load devices, had a "normal braking ratio of 1." He further testified that at the time of the train derailment, the operating rules had not been changed to reflect this. The DRGW cars were not equipped with EL devices. All 69 hopper cars were equipped with composition brake shoes.

Following the derailment, many wheels and brake heads were inspected. This was a random inspection of available parts because many parts were buried and almost none of the parts could be identified as belonging to any particular car or part of the train. Of a possible 552 brake heads on the train, 160 were examined with the following conditions noted: 36 had been burned away, 102 showed signs of heavy heat and excessive braking, and 22 showed light or no signs of excessive braking although most of these showed signs of service wear. According to SP’s chief mechanical officer, some showed no signs of heavy braking because of "…the variation in the brake shoe thickness, the thickness of the wheels…and braking forces. They are not exactly the same on all cars." He further testified that braking forces are not evenly distributed even on one car. Of a possible 276 wheel sets, 142 were inspected of which 109 showed obvious evidence of overheating from heat buildup by excessive or heaving braking. The chief mechanical officer testified that based on SP’s postaccident inspection of the wheels and brake heads, he believed that the brakes on Extra 7551 East were effective and that the brake pipe was intact.

Locomotive wheels and brake shoes showed heaving braking and heat on every unit. Some brake shoes had been burned away and the backing plate had begun to melt.

Mechanical Information

Use of Dynamic Brakes.—According to the Associirtion of American Railroads’ Director of Safety and Operating Rules, many Class I railroads emphasize the use of dynamic brakes to control a train, thereby conserving fuel and minimizing brake shoe wear. Rule 58F of the SP Air Brake Rules and Train Handling Instructions states, "The dynamic brake must be used whenever practicable in reducing and controlling train speed.…" Rule 58I further states, "Where the available dynamic brake will not properly control the speed of the train, the automatic air brakes must (then) be used to an extent which will allow the dynamic brakes to be reduced to a value where it will be flexible enough to control changes made in speed due to physical characteristics of the road." The Safety Board is aware that similar rules exist on other railroads. Rule 501B of the Burlington Northern Air Brake and Train Handling Rules states:

Train handling must be performed in a manner that will be most fuel efficient consistent with good train handling. Therefore, maximum use must be made of the trottle modulation, throttle reduction and dynamic braking methods for slowing, controlling, and stopping trains. Unless rules specify otherwise, DURING PLANNED BRAKING OPERATION, IF ONE OR MORE OPERABLE DYNMIC BRAKES ARE AVAILABLE THE POWER BRAKING METHOD WILL NOT BE USED."

Of SP’s road fleet of 2,100 units, 1,800 units, according to the chief mechanical officer, are equipped with dynamic brakes. SP locomotives are designed such that when the train brakes are applied in emergency, an interlock will nullify the dynamic braking. According to SP’s chief mechanical officer, the system is designed in this manner "…to prevent train handling problems in the case of a break in two [a separation of two cars] and to prevent wheel slide because of excessive braking which would be the combination of the electric [dynamic] braking and the independent brake.…" He could offer no explanation as to why some railroads have modified the system to retain dynamic braking when the train brakes are applied in emergency. He stated that the SP had checked with the manufacturer and that the manufacturer "…will not make that modification for the SP or any other railroad." He further stated that the SP was not considering modifying the locomotives. The Safety Board contacted one manufacturer who indicated that any specifications requested by a carrier, as long as they were in compliance with Federal regulations, would be made. The Safety Board is aware that the Union Pacific and the Burlington Northern have their own retrofit program to eliminate the interlock feature.

Maintenance Reports and Reporting of Defective Locomotive Units.—SP Rule 2A requires the engineer to report locomotive defects to the dispatcher and to fill out a form outlining the defects. This form remains in the locomotive cab until the locomotive reaches an appropriate facility where mechanical department personnel can make the repairs. The head-end engineer testified that he complied with both parts of this rule with respect to the inoperative dynamic brakes on the lead locomotive unit, 7551. The helper engineer testified that he did not inform the dispatcher that the dynamic brakes on one of his helper units were inoperative because the dynamic brakes were inoperative when he began his tour of duty and he believed that the engineer whom he had relieved had reported the defect to the dispatcher. The assistant chief dispatcher who assigned the power (locomotive units) for the movement of Extra 7551 East testified that he does not request information from engineers and that he does not query the computer system^[24] about the status of dynamic brakes on locomotive units. He further testified that it is the responsibility of engineers to inform him of any locomotive defects. He also stated that there are no written procedures that specifically address what to do with information received from engineers regarding defective locomotive equipment.

The chief mechanical officer testified that engineers, in addition to reporting defects to the dispatcher and filling out the appropriate form, will occasionally report defects directly to the roundhouse (engine repair facility) foreman. He further testified that mechanical department personnel, if they become aware of any defects, will update the computer with information. According to the chief mechanical officer, the dispatcher, once he receives information from engineers regarding defects, has the responsibility to update the computer. The assistant chief dispatcher testified that he often updates the computer when he receives reports of defects, although he believed it was not his responsibility to do so, or he will give the information to a clerk who will then update the computer when time is available.

A review of maintenance records and failure reports by Safety Board investigators revealed that on May 4, an engineer had filed a failure report on unit 7549, the third unit in the head-end consist, noting that there were no dynamic brakes. According to the maintenance record, the motor braking switch was stuck and the repair was made. The chief mechanical officer testified it was not a major repair and that there was a possibility that a defect of that type could occur again. With respect to unit 9340, the fourth unit in the head-end consist and the one that the head-end engineer reported as "intermittent' in dynamic braking, maintenance records indicated that it had received extensive repairs to the dynamic brake on April 27 and 29, 1989. According to the chief mechanical officer, the extensive repairs would indicate to him that the dynamic brakes should have been working on the day of the accident. According to the failure reports, unit 8317, the lead unit in the helper consist, had been reported as having inoperative dynamic brakes on May 8, 1989, 4 days before the accident. The chief mechanical officer testified that it was not uncommon for a unit to continue to be used in helper service "until it worked its way" to the Los Angeles repair facilities. Testimony by the head-end engineer and the helper engineer indicated that it was not uncommon to have a unit in a locomotive consist with inoperative dynamic brakes. The chief mechanical officer testified that the number of units reported to have inoperative dynamic brakes varied on a daily basis from 3 to 35.

Recovering Dynamic Brakes.—According to the chief mechanical officer, an engineer can recover the dynamic brakes (after an emergency application of the train brakes has been made) by going to "a handle off position and recover[ing] the PC after about 70 seconds."^[25] He stated that he believed the head-end engineer had sufficient time to recover his dynamic brakes. He also stated, "I suspect there could have been some slight benefit going back into dynamic brakes but at those speeds the dynamic braking effort is very, very low."

Federal Railroad Administration’s (FRA) Position Regarding Functioning Dynamic Brarkes.—The Safety Board received conflicting testimony from SP personnel regarding the company’s interpretation of FRA requirements for functioning dynamic brakes. The general road foreman of engines stated that he believed, based on his interpretation of FRA regulations, that if a locomotive unit is equipped with dynamic brakes, they "must operate." The chief mechanical officer stated that he believed there were no FRA requirements regarding functioning dynamic brakes. The Safety Board requested the FRA to provide in writing its position on functioning dynamic brakes. In a letter to the Safety Board dated October 18, 1989 (see appendix G), the FRA stated:

The Railroad Power Brake and Drawbars Regulations does [sic] not require the presence of a dynamic brake. However, dynamic brakes are referred to the in the Locomotive Safety Standards, which states in part "If a dynamic brake or regenerative brake system is in use, that portion of the system in use shall respond to control from the cab of the controlling locomotive."

This part makes clear that both the equipping and the use of dynamic brake is optional. The FRA will not take exception if a dynamic brake is found inoperative or operates at less than maximum designed capacity.

Southern Pacific’s Method of Operation

Air Brake Rules and Timetable Instructions.—Trains operating on the double main track over the Mojave Subdivision are controlled by the train dispatcher using Direct Traffic Control between Mojave and East Mojave. Between Ansel and West Colton, trains are operated in accordance with signal indications of an automatic block and traffic control system.

Timetable No. 2, effective October 25, 1987, was current at the time of the accident. Maximum allowable speed on the line between East Mojave and West Colton was 65 mph for freight trains. Exceptions to the maximum allowable speed for eastward freight trains between East Mojave and West Colton were as follows:

between MP 463.8 and MP 487.4⁠	30 mph
between MP 487.4 and MP 491	40 mph
between MP 491 and 491.9	30 mph
between MP 491.9 and 492.7	15 mph

The SP had adopted the General Code of Operating Rules, which became effective on October 28, 1985. The SP’s Air Brake Rules and Train Handling Instructions, last revised on November 1, 1985, were also in effect. Pertinent excerpts from the Air Brake Rules and Train Handling Instructions follow:

Rule 2. Dynamic Brake.

Helper locomotives entrained may not use dynamic brake unless road engine has operative dynamic brake.

The number of axles of dynamic brakes of the helper engine(s) will be added to axles of dynamic braking of the road engine to determine the tonnage that may be handles in accordance with applicable Air Brake Rules.

Dynamic brake on head end of freight trains must not exceed 24 axles. Each helper entrained must not exceed 36 axles.

Rule 17. Retaining Valves.^[26]

The Superintendent will prescribe the number and locations where retaining valves must be used.^[27]

Instructions in Timetable No. 2 indicate that for the descending grade between Hiland and West Colton, retaining valves will be used under certain conditions. For trains being operated down the grade without operative dynamic brakes, one retaining valve will be used for each 80 tons in train. If gross tonnage exceeds 80 tons per operative brake, retaining valves must be used on all cars and speed must not exceed 15 mph. For trains being operated with operative dynamic brakes, use of retainers is not required if tons per axle of dynamic brake does not exceed 375 per standard range or 450 per extended range.

Rule 33. Tonnage Per Operative Brake.^[28]

The maximum tonnage per operative brake that may be handled on descending grades of 1.8 percent or over will be prescribed by the superintendent.

Freight trains handling cars with single capacity brakes (*), with tonnage exceeding 80 tons per operative brake, must not exceed 45 mph, except maximum speed must not exceed: (1) 25 mph; or (2) 20 mph in grade territories as designated by Superintendent by milepost locations under appropriate subdivision.

*Loaded cars with empty-load brakes are to be considered the equivalent of one and one-half (1½) cars in determining tons per operative brake.^[29]

Tonnage of operating locomotive(s) not in dynamic braking is not to be used in determining tons per operative brake.

The instructions in Timetable No. 2 indicate that the descending grade between Hiland and West Colton is covered by rule 33. The timetable also lists the maximum tons per operative brake for trains descending the grade and the exceptions for those trains using dynamic braking (appendix H). The instructions also state:

Insufficient dynamic brake capacity or failure of dynamic brake which results in exceeding these tonnages per axle, is to be considered as operating without dynamic brake.

Should dynamic brake failure occur on one or more locomotives resulting in insufficient dynamic brake capacity, train must stop and all retaining valves turned up. Train may then proceed not exceeding 15 mph if, in the judgement of the conductor and engineer, it is safe to do so.

The SP’s general road foreman of engines provided the Safety Board with a speed decision flow chart for Rule 33 (see figure 15). According to his testimony, "A train consisting of 69 cars with a weight of 8,900 tons and that had 18 operative dynamic brake axles" would not have been allowed to descend the grade east of Hiland. Extra 7551 East on the day of the accident had 128 tons per operative brake (8,900 trailing tons divided by 69 (number of cars in train, not using the 1½ braking equivalence)) and 494 tons per axle of dynamic braking (8,900 trailing tons divided by 18 (three locomotive units with six axles each)). Using the speed decision flow chart, the general foreman illustrated why the train was not permitted to operate (follow arrow #1 on figure 15). Using the chart, the general foreman also illustrated the decision process the engineer would have made on the day of the derailment with the information that he had 69 tons per operative brake^[30] (follow arrow #2 on figure 15). According to the general road foreman, "If the train would have 6,151 tons, with the information that [the head-end

Figure 15.—Speed decision flow chart for Rule 33.

engineer] had and the dynamic brakes [he] thought he had working, he could easily have controlled the train down the hill." He further stated that the engineer, based on the information provided to him, could have taken the train down the hill without any dynamic brakes. According to the head-end engineer, based on the information he had, rule 33 did not apply to his train.

As outlined in the Air Brake Rules and Train Handling Instructions, the dynamic brake retarding force per brake axle diminishes as speed increases. For example, at a speed of 23 mph, the dynamic brake retarding force per axle is 10,000 lbs; at a speed of 40 mph, the dynamic brake retarding force per axle is 5,700 lbs.

Rule 61.E. Balancing the Grade

Operating freight trains on descending grades involves:

1. Balancing the grade, or holding speed steady at safe and practical values.

The amount of brake (train) retarding force used to balance the grade normally should not exceed one half (50 percent) of the normal full service train brake available if dynamic brake and pressure maintaining are operative.

In order to hold speed steady on a descending grade, the force of gravity must be balanced by the sum of train resistance and brake retarding force. The heavier the grade, the lower the effect of train resistance; and the more brake must be used. Train resistance will vary with the type of cars, train make-up, and train length and weather. On heavier grades the majority of the grade retarding force comes from the dynamic brake and the train air brake.

The locomotive engineer, the helper engineer, the road foreman of engines, and the general road foreman all testified that they considered rule 61.E.1 to be a recommended general guideline or an option rather than a requirement. Testimony also indicates that engineers have routinely gone beyond the 50 percent reduction. On May 17, 1989, SP issued train order No. 1903, adding the following new rule to their operating rules:

Rule 627.8.

Within the territories where air brake rule 33 applies, except on Yuma subdivision–Los Angeles division, and with the use of dynamic brake the following brake pipe reductions must not be exceeded to control the train at the following speeds:

Maximum Speed	Maximum Air Brake Pipe Reduction
30 mph	13 pounds
25 mph	16 pounds
20 mph	18 pounds

In the event train speed cannot be controlled without exceeding the above brake pipe reductions, train must be stopped, secured and air brake system recharged. Train must not proceed unless authorized by the chief train dispatcher.

According to the general road foreman of engines, the SP decided to "put definite limits on what [speed] a train could go with a certain air brake reduction to reduce the wheel heat and keep it within the limits." He further stated that the Rio Grande had conducted tests and determined that an 18-pound reduction at 20 mph and a 13-pound reduction at 30 mph could be made without excessive wheel heat.

By special instructions, effective May 31, 1989, speed restrictions were placed on the area from Hiland to West Colton (the descending grade). According to the special instructions, trains with 25 or more loads of coal, grain and/or bulk minerals must not exceed 20 mph.

Rule 13 of the SP Air Brake and Train Handling Rules addresses the procedure for placing the locomotives in reverse. The rule states, "Should it become impossible to stop a train with the air brakes…place throttle in IDLE position, apply sand, place reverser lever in the opposite position and move the throttle to No. 1 position."

Communication Between Head-end and Helper Engineers.—On the day of the train derailment, there were no requirements that the head-end engineer and helper engineer communicate with each other regarding the condition of their respective locomotive units. Both the road foreman of engines and the general road foreman of engines testified that based on their review of the radio transcripts between the head-end engineer and the helper engineer on the day of the derailment, the amount of communication that took place was less than what they would have expected. The helper engineer testified that he communicates with the head-end engineer by observing the air gauge. According to his testimony, he can determine what actions the head-end engineer is taking by observing the air brake reductions.

Effective May 22, 1989, by special instructions, the following rule was added to the Western Region:

The road and helper engineer(s) must communicate the condition of their units and train in order to determine maximum authorized speed and train handling requirements. Helper engineer will observe speed indicator while running and remind road engineer of speed requirements if necessary. If helper engineer is unable to communicate with road engineer and if train continues to operate in excess of maximum allowable speed, helper engineer will take necessary action to stop train.

Tonnage Information for Cars.—At the time of the train derailment and when yard clerks at various outlying areas released a car as loaded, SP’s computer system required that certain information be entered into the system

The director of clerical operations testified that the clerks in the various outlying areas are responsible for checking the accuracy and completeness of shipper-tendered bills of lading. According to his testimony, the first-line supervisor for these clerks is located in Los Angeles. He further stated that during the last few years, shippers have been sending their bill of lading information directly to the central office in Los Angeles rather than dealing with clerks at the various outlying areas.

The Calnev Pipeline

Description

The Calnev pipeline was constructed by Mid-Mountain Contractors, Inc., during 1969 and 1970. The approximately 248-mile-long pipeline, which transports petroleum products including gasolines, jet fuels, and No. 2 diesel fuel, originates at the Colton Pump Station at Colton, California, and terminates at Las Vegas, Nevada. From the Colton Pump Station (elevation 1,040 feet), the pipeline route is generally northward and crosses Cajon Pass at an elevation of 4,480 feet at MP 28 (figure 5). From Colton to about MP 236, the pipeline is 14 inches in diameter, and from MP 236 to the Las Vegas terminal, the pipeline is 8 inches in diameter. The first 107-mile section of the 14-inch-diameter pipeline was constructed of the same grade of pipe that was involved at the rupture site. The pipe at the rupture site was manufactured of steel by Kaiser Steel Corporation to American Petroleum Institute standard 5LX 52, using an electric resistance welding process. The pipe had a 0.312-inch wall thickness and weighed 45.61 pounds per foot. As a minimum, the pipe was required to have a specified yield strength of 52,000 psi and a specified tensile strength of 66,000 psi. Records of tests performed on the steel used to manufacture the pipe indicates that the steel exceeded these minimum requirements with some tests showing minimum specified yield strengths of 66,000 psi and minimum specified tensile strengths of 74,430 psi or greater. The pipe was coated with TGF3, a coal tar base coating. According to Calnev's cathodic protection records, the pipe had a minimum negative (cathodic) voltage of 0.85 volts (generally it had a considerably more negative voltage) as measured between the pipe and the soil. A cathodic protection rectifier was located at the Colton Pump Station, and Calnev's records indicate that there had never been a corrosion leak found on this 14-inch pipeline system. Calnev's manager of operations testified that if the coating damage existed prior to the derailment, Calnev would not have been able to see any change in the cathodic protection in this case because, "There is a casing that runs under Highland Avenue. At this particular location the casing and the pipe are operating at the same potential. That large casing would probably mask any damage to the coating that might be evident in that location. I don't think you would have seen a change to the cathodic level there."

The first 107 miles of the pipeline were hydrostatically tested between June 20, 1970 and July 3, 1970; the section through the rupture site (MP 0.0 through MP 25.2) was tested on June 29 and 30, 1970. The pressure test on this section was begun at 8:15 a.m. on June 29, 1970, at 2,085 psig and completed at 12:30 p.m. on June 30, 1970, at 2,083 psig. The minimum pressure during the test was 2,075 psig, and the minimum 4-hour internal sustained pressure was 2,077 psig.

Check Valves

At the time the pipeline was constructed, Calnev installed check valves in its pipeline to prevent backflow of product from one section of the pipeline to another. These valves also serve to minimize the amount of product that can be released from the pipeline should the pipeline rupture. Generally, Calnev installed top-hinged check valves, and at some locations there are connections installed to bypass the check valves. However, on the 14-inch portion of the pipeline, Calnev installed seven Wheatley "All-Clear Check Valves." These check valves are side-hinged check valves which purportedly provided advantages over the top-hinged check valves by producing less pressure drop and offering less resistance to the passage of spheres and scrapers. Side-hinged check valves were installed at MP 0.0, 6.9, 14.9, 19.2, and 25.7. Calnev's manager of operations testified that he was not aware that Calnev had ever inspected any of the check valves installed in the pipeline between the Colton pump station and Cajon Pass to determine if the valves operated properly. He further testified that it was his understanding that check valves are not routinely inspected in the industry and that he was unaware of any Federal regulation or industry standard that required such inspection. He stated that following the rupture Calnev made plans to inspect the check valves in this area. In a letter to the Safety Board dated May 21, 1990, Calnev stated, "Calnev has installed four new check valves;[sic] three to replace existing check valves and one additional check valve at MP 10.0. Our intention is to replace one more check valve and install a supplemental block valve near another in the next 60 to 90 days."

The OPS representative who testified at the Safety Board's public hearing stated that the proper operation of check valves can be important to the safe operation of pipelines; he also advised that the OPS historically has considered that the regulations do not apply to the maintenance of check valves. The OPS has not issued an interpretation to this effect and it has not provided to its enforcement personnel any guidance indicating that check valves do not have to comply with the maintenance requirements; however, the OPS representative stated that this position reflected what OPS has been doing from an enforcement policy.

The Calnev manager of operations further testified that, based on the amount of product eventually required to refill the line, at the time of the rupture, the check valve at MP 6.9 did not close, the valve at MP 14.9 "must have come closed at some point," and that check valve [sic] at MP 19.2 "probably has at minimum leaking seats."

Block Valves

Remotely operated block valves were installed on the Calnev pipeline at MP 35.4 and MP 46.7. A manually operated block valve was installed at MP 25.7. According to the testimony of the incident commander (the deputy fire chief) and Calnev's manager of operations, the deputy fire chief requested after the train derailment that a block valve be installed just north of where the derailment occurred. According to Calnev's manager of operations, "With a block valve you have the ability for positive shut-off. You can turn a crank and tighten it and possibly have a more certain measure that your pipeline is shut off at that point. I think the chief felt that given the difficulty we had in getting that check valve to seat during our drain-down, that that might be a good idea given the population in the area…We were basically in agreement with the chief that that would be a good idea." He further stated, "There is a fair amount of lead time in ordering such an item and a fair amount of time to set up an installation such as that one." Subsequent to the pipeline rupture, a remotely operated block valve was installed at MP 6.9.

Dispatch Center

The pipeline system is controlled by dispatchers from a dispatch center at the Colton Pump station. The system is equipped with a monitoring system that scans selected system parameters, such as pipe pressures and motor drive amperages, every 13 seconds, compares the data with programmed acceptance values, and through visual and audible alarms, alerts the dispatcher to changes to operating conditions in the system and abnormal or unacceptable occurrences. The audible alarm indicates that a change has occurred; however, this does not necessarily indicate that there is an emergency or that any action is required on the part of the dispatcher other than to acknowledge the alarm by pressing a key on his terminal keyboard. The visual alarms are presented in the form of numerical values flashing on a colored background. The background color varies depending on the measured value for the particular operating parameter. Background colors range from shades of white and blue, representative of the range of low pressure conditions, to yellow and red, representative of the range of high pressure conditions. Normal ranges are presented on a green background.

A computer printout of the monitoring system indicated that on the day of the accident, the dispatcher on duty received both a low suction and a low discharge pressure alarm on his computer terminal screen. The dispatcher did not detect the low discharge pressure alarm, and by one stroke on his terminal keyboard, he silenced the audible alarm, caused the flashing word "alarm" to disappear from his screen, and caused the flashing numerical information regarding the low suction pressure and the low discharge pressure to return to a steady presentation; the background color does not change until the operating condition changes. According to Calnev, subsequent to the rupture, Calnev modified its automated control system to include a high flow set point whereby if excessive flow is observed out of the Colton pump station (indicative of a potential leak or rupture), the system will automatically shut down the Colton pump, and indicate the alarm condition.

Emergency Response Manual

On the day of the pipeline rupture, Calnev did not have any procedures in its abnormal operation response plans (a section of the company's emergency response manual) that would advise the dispatchers of the actions to take upon receiving both a low discharge pressure and a low suction pressure alarm. Calnev's manager of operations stated, "We felt that it was adequately covered in the explanation section for low suction pressure" which advises that the line pressure be checked in the event of a low suction pressure alarm. He stated further that following the pipeline rupture, Calnev revised its manual to include an explanation of what to do in the event a low discharge pressure alarm is received.

Calnev's emergency response manual was last revised in January 1989. This manual contains a list, by milepost, of telephone numbers for fire and police departments, and procedures for notifying Calnev personnel and other agencies in the event of a spill or leak. The manual also contains maps of the pipeline and directions to each mainline block valve, and procedures for responding to a natural disaster and external incidents.

The procedures for a suspected leak require the pipeline to be shut down, pressures to be stabilized, remotely operated valves to be closed, and pressures in specific sections of the pipeline system to be monitored. If a leak is confirmed, the procedures outline specific actions to be taken to locate the leak and to respond to the leak.

The procedures for a natural disaster and external incident refer to the potential adverse effects of train derailments. The procedures indicate that substantial portions of the pipeline system are built on the railroad right-of-way and that train derailments pose a serious threat to the pipeline primarily by equipment being used to clear the wreckage and replace the roadbed. The areas where the pipeline system is located near railroad tracks are listed by milepost; the area of the train derailment of May 12, 1989, is included in this section. In the event of a train derailment, the procedures indicate that Calnev personnel are to be immediately dispatched to the scene and assess the situation to determine if the pipeline could have been damaged. Railroad personnel are to be contacted and advised of the location of the pipeline. In the event of possible damage, the pipeline is to be shut down, and upstream and downstream valves are to be closed. The procedures also indicate that once the pipeline has been secured, the location of the pipeline through the derailment area should be accurately marked, heavy equipment should not be allowed to operate over the pipeline if it is considered hazardous to the pipeline, and Calnev personnel should be present on scene until all work is completed.

Personnel Information

Operating Crew of Extra 7551 East

The head-end engineer had been off duty for about 20 hours before reporting for duty at Bakersfield at 9:00 p.m. on May 11. The engineer reported the following information: He spent his off-duty time sleeping, eating, watching television, and relaxing. He had been eating regular meals during the day preceding the accident, had been receiving his usual amount of rest of about 10 hours, and was fully rested when he reported for duty on the evening of May 11. There had been no recent changes in his lifestyle, he had not consumed any alcohol during the days preceding the accident, and he was not a user of illicit substances.

The engineer held an active State of California driver's permit. An inquiry to the State of California Department of Motor Vehicles (SCDMV) revealed that the engineer had no history of having received any summons or convictions. The National Driver Register (NDR) contained no information on revocations or suspensions regarding the engineer's driving privileges.

The head-end engineer had been employed by the SP for almost 15 years at the time of the accident. He had held the positions of fuel oil attendant, laborer, and fireman before being promoted to the position of engineer on November 28, 1986. (For additional information, see Engineer Training Program.)

The head-end engineer had been qualified on the physical characteristics of the territory by making one check ride from Techachapi to Bakersfield (see figure 1) with a supervisor in September 1988. He stated that he was familiar with the descending grade in the accident area and had operated trains over this trackage several times. He stated further that he had previously operated trains with a trailing tonnage of 6,151 tons and with a trailing tonnage of about 8,900 tons. His testimony also indicated that he had never operated a train that he believed the tonnage of which was substantially more than the tonnage shown on his train documents. He did indicate, however, that this was the first unit (single commodity) freight train he had operated through the Cajon Pass; all of his prior experience through the Pass was operating mixed commodity freight trains. He added that he believed this was the first time he had transported trona. The head-end engineer stated that he had worked previously with the other head-end crewmembers [sic], but had no knowledge of, nor had previously worked with, the helper engineer.

The conductor of Extra 7551 East had been off duty the 4 days preceding the accident. The conductor's wife reported the following information about the conductor: On Thursday, May 11, the conductor awoke around 8:30 a.m. and remained at home during the day. He received his call for duty, as expected, and reported to the Bakersfield yard at 9:00 p.m. that evening. He had been eating regular meals and had been receiving his usual amount of rest during the days preceding the incident. Her husband was "rested as usual" when he reported for duty the evening of May 11. She had noticed no changes in her husband's lifestyle. The conductor did not smoke cigarettes or drink alcohol.

The conductor held an active State of California driver's permit. According to the SCDMV, the conductor had no history of having received a summons or conviction. The NDR contained no information on revocations or suspensions regarding the conductor's driving privileges.

The conductor had been employed by the SP for 17 years at the time of the accident. He had held the position of brakeman until April 15, 1975, when he was promoted to the position of conductor.

The head-end brakeman of Extra 7551 East had been off duty during the 48 hours preceding the accident. The brakeman's wife reported the following information about the brakeman: He spent the time during the days conducting personal business and engaged in activities with his family. On Thursday, May 11, he awoke about 9:30 a.m. having received about 10 1/2 hours of sleep, and spent the day at home. He reported for duty at Bakersfield at 9:00 p.m. that evening. He had been eating regular meals, had been receiving his normal amount of rest, and "was not fatigued" when he departed home on the evening of May 11. He did not smoke cigarettes, drink alcohol, or use illicit substances, and she had not noticed any recent changes in her husband's lifestyle.

The head-end brakeman had been employed with the SP for more than 17 years at the time of the accident. He was promoted to the position of brakeman on November 27, 1971.

The helper engineer had been off duty since 11:00 p.m., May 9, having completed at that time an approximate 10-hour tour of duty. He stated that on May 10, he attended a union meeting in the morning and for the remainder of the day engaged in personal activities. According to his testimony, on Thursday, May 11, he awoke around 10:00 a.m., having received about 8 hours of sleep. He spent the day performing personal business and retired that evening about 11:30 p.m., at which time he received a call from the crew dispatcher for a 1:30 a.m. duty call. He reported to the West Colton yard and then rode in a company van for the 1/2-hour trip to the Dike siding where he was to relieve the on-duty helper engineer.

The helper engineer reported that there had been no recent changes in his lifestyle, that he does not use illicit substances, and that he had not consumed any alcohol during the days preceding the accident.

The helper engineer stated that he had eaten regular meals during the days preceding the accident and that he normally receives 6 to 8 hours of sleep daily. In his initial statements to Safety Board investigators, he stated that when he received the call for duty on the evening of May 11, he had not received his proper rest and "was tired." He elaborated by stating that he was not tired when he first reported for duty but that he was not "in tip top condition the whole trip." When questioned if he had fallen asleep during the trip, the helper engineer replied, "I don't think so." The engineer further stated that he had expected to receive a call for duty because he had called the crew dispatcher's office several times that day, but believed that he would receive the call for duty later in the night or early the following morning. During the Safety Board's public hearing, he testified that he was not tired when he reported for duty and had no difficulty remaining alert during the trip.

The helper engineer had been employed by the SP for more than 11 years at the time of the accident. He had held the positions of hostler and fireman before being promoted to the position of engineer on November 5, 1979.

The helper engineer stated that he normally operated trains between West Colton and Yuma. He was not qualified on the physical characteristics of the railroad for the territory in which the accident occurred and could not, therefore, operate as a road engineer in this area. He estimated that during the past year he had served as a helper engineer about four times on trains operating over the accident territory. Company records indicated that during the month preceding the accident, the helper engineer had not worked with any of the other crewmembers [sic] assigned to the accident train.

The helper brakeman received an emergency call for duty from the crew dispatcher on the evening of May 11, to report for duty at 1:30 a.m., May 12. He stated that he had expected to be called for duty about 10:00 a.m. later that morning. Prior to the emergency call, the brakeman had been off duty since 9:00 p.m. on May 10. The helper brakeman reported the following information about his activities: He had "a normal day" on May 11, had been eating regularly (which for him was one meal in the evening) during the day preceding the accident, had been receiving his usual amount of rest, about 8 hours daily, and he was not fatigued when he reported for duty on the day of the accident. He had consumed one beer at home on the evening of May 11. His lifestyle had been altered as a result of his wife's death 3 months earlier. He did not indicate that he was not adjusting properly to this loss.

The helper brakeman had been employed by the SP for more than 38 years at the time of the accident, holding the position of brakeman since the time he was hired. He estimated that he had been a crewmember [sic] on trains operating over the accident area on about 10 occasions in the past and that he had worked on many occasions with the helper engineer.

On-scene investigators attempted to locate the grips (personal bags) belonging to all five crewmembers [sic]. It was learned that the helper crew had taken their bags when they departed the accident site. The grip belonging to the conductor was removed by the wreckage by railroad officials, and investigators were unable to locate any documentation concerning the contents of this grip. The grips belonging to the head-end engineer and head-end brakeman were located in the wreckage and recovered. A review of the contents of these grips revealed nothing noteworthy.

Other Southern Pacific Personnel

The train dispatcher on duty at the time of the accident normally worked a 5-day week. Prior to the day of the accident, the dispatcher had not worked since May 6, due to illness. She stated that she was feeling fine when she reported for duty on the morning of May 12. The dispatcher had been employed by the SP for almost 10 years and had held positions as yard clerk and interlocking operator before being promoted to the position of dispatcher on November 19, 1988.

The assistant chief dispatcher, who arranged the locomotive units for the movement of Extra 7551 East, worked a regular shift of 10:30 p.m. to 6:30 a.m., 5 days a week. He had been off duty for 16 hours before reporting for duty on the evening of May 11. The assistant chief dispatcher was employed by the SP in July of 1970. He held various positions including freight clerk, yard clerk, and train order operator until being promoted to the position of train dispatcher in 1973. He was promoted to chief train dispatcher in August 1976, resigned voluntarily from that position in September 1977, and returned to the position of train dispatcher in Los Angeles until April 1983. At that time, he exercised his seniority options and returned to Bakersfield as a crew dispatcher and worked in that capacity until 1985, when he returned to the train dispatcher position. His last examination on the operating rules was conducted in 1985.

Calnev Pipeline Dispatcher

The dispatcher on duty at the time of the pipeline rupture had been employed with the Calnev Pipe Line Company since October 3, 1988. He was hired as a pipeline operator, which includes serving as a relief dispatcher. He was performing the duties of relief dispatcher at the time of the accident.

According to the dispatcher, the day of the pipeline rupture was the third day of his work week; he had finished his last shift at 3:00 p.m. the preceding day. On the day of the rupture, he reported for work at 6:45 a.m. He reported the following information: He had been receiving his usual amount of rest and was properly rested when he reported for duty. He was not taking any medication on the day of the pipeline rupture, had not consumed alcohol the day before the rupture, and he does not "involve himself" with illicit drugs.

The dispatcher had been employed previously with the Paramount Petroleum Corporation for 10 years, during which time he served as a pumper-pipeline operator, a laboratory technician, and a crude oil unit operator.

(Additional personnel information is in Appendix B.)

Southern Pacific Training Programs

Engineer Training Program

Trainees for the engineer training program were selected from employee applications with preferential treatment given to those applications submitted by United Transportation Union (UTU) members—brakemen, switchmen, and hostlers—because of existing labor agreements between the SP and the UTU. Those trainees selected initially entered a 4-week formal training program during which preliminary air brake, mechanical, locomotive, and operating rules are covered both in the classroom and in the field. The class size for the program normally consisted of 10 trainees. If the trainees successfully completed examinations midway and at the end of the 4-week period, then they progressed to the next stage, which consisted of making 60 road trips with a qualified engineer. A trainee was not assigned to a specific engineer during this time (labor agreements did not provide for instructor engineers), and, thus, may have ridden with many different engineers in the process of completing 60 road trips. Following the completion of 60 road trips, the trainees were evaluated by the road foreman of engines on the respective district over which they had been working. If he determined that the trainees had reached a minimum level of proficiency, then they were scheduled for the final 3-week phase of training at the company's training facility in Cerritos, California: 1 week consisted of 40 hours of classroom instruction; the last 2 weeks consisted of 1/2 day of classroom instruction and 1/2 day of simulator training. If the trainees successfully passed all three written examinations (one each on air brakes, mechanical systems, and operating rules) and demonstrated train handling skills as observed in the train simulator, they were then promoted to the position of engineer and received a seniority date. An engineer was not qualified for a given territory until the road foreman of engines for the territory had ridden with the engineer for a period of time and had determined that the engineer was knowledgeable of the territory and could adequately handle trains over the territory. (According to the assistant manager of training and development, the number of times a road foreman of engines would ride with an engineer varied based on the level of skills of the engineer.)

The SP also had in place a 1-week and a 2-week continuing education program during which time engineers returned to the Cerritos facility for refresher training. The 1-week program consisted primarily of reviewing train handling skills (1/2 day in the classroom and 1/2 day in the simulator) and was geared for engineers who worked in heavy-grade territory or mountainous terrain. During the 2-week program, train handling skills were reviewed, and the mechanical systems on the locomotive and the operating rules book were also reviewed. The engineers were not confronted with a pass/fail situation upon completing the continuing education programs. The superintendent of an engineer's respective division received a report on the engineer's performance both on the simulator and on the written examinations. The superintendent could then use the information to determine if the road foreman of engines should spend additional time with a particular engineer.

The head-end engineer of Extra 7551 East entered the engineer's training program on October 20, 1986. After successfully completing the 2-week classroom or "presimulator" training course, he attended the 3-week training course held at the training center in Cerritos. After successfully completing 1 week of classroom instruction and 2 weeks of simulator training at the center, he was promoted to the position of engineer on November 28, 1986. The engineer returned to the training center in January 1988 for the 1-week continuing education program to receive additional instruction on heavy-grade operations. The engineer successfully completed both the classroom portion and the simulator training portion of the program.

The head-end engineer of Extra 7551 East testified that he had never been trained on procedures concerning the reversing of engines, had never received instruction concerning the effects of extended brake application on the deterioration of brake shoes, had never received instruction regarding train handling while receiving helper engine assistance, and had never been placed in an emergency situation during simulator training. He further stated that he was not taught during training how to recover dynamic brakes after an emergency application of the train brakes had been made.

The helper engineer entered the engineer's training program on August 13, 1979. He successfully completed the final phase, 1 week of classroom instruction and 2 weeks of simulator training, before being promoted to the position of engineer on November 5, 1979. He returned to the training center in Cerritos in July 1988 and successfully completed a 2-week continuing education program. The helper engineer testified that during his training, the company rule that addressed reversing the engines was discussed in situations involving "light engines or just a couple of cars, low speeds." He further testified that during this simulator training, they operated trains with helper units. He stated, "...you are trained to take and just go by what the road engineer requests. Normally, it is standard procedure just to go in full dynamics, unless he requests otherwise, and stay in full dynamics."

According to SP's assistant manager for training of engineers, reversing the engines was not taught during any aspect of the training program "because with the train in emergency, we do not allow the engineer to attempt to reset the PC switch before the train comes to a halt." His testimony also indicated that emergency situations incorporated into the simulator training were predicated on the premise that once the brakes are applied in emergency, the train will stop. With respect to helper engine service, the assistant manager for training stated, "The extent of our instruction to people as far as being helper engineers is push as hard as you can up the hill and hold back as hard as you can going down the hill and if the road engineer asks you to do something, do it."

Dispatcher Training Program

The SP was training its dispatchers at its training center in Cerritos. According to the training officer for dispatchers, the existing program had been in place for about 1 1/2 years. Candidates for the dispatcher position entered an 8-week training course that incorporated the use of the same computerized dispatching equipment that the individual would use once assigned to an office. After passing the final examination on the classroom portion of training, candidates were sent to a dispatching office where they began their on-the-job training. There was no set period of time that trainees were required to perform on-the-job training. The chief train dispatcher determined when an individual was qualified for a particular dispatcher's position.

The dispatcher, who had operational responsibility over the Mojave Subdivision and was on duty at the time of the derailment, successfully completed the 8-week dispatcher training program on August 19, 1988. She then received on-the-job instruction from an experienced dispatcher for 3 months before being qualified to operate independently as a dispatcher. The assistant chief dispatcher, who assigned the locomotive units for the movement of Extra 7551 East, had not been through the Cerritos dispatcher training program; his training for the position of dispatcher was all on-the-job training.

Clerk Training Program

The yard clerks who estimated the weight of the cars at the time the cars were released and the yard clerk who estimated the weight of the trona on the shipper's bill of lading had received no formal instructions regarding their duties, according to their testimony. All training had been on-the-job training with other clerks. According to the director of system clerical operations, "It's not always feasible to give these people classroom training when, in fact, they may be trained in a classroom for 2 weeks and then have somebody exercise their seniority against them or they bid to another position...." He estimated that about 20 percent of the clerks were receiving classroom instruction and that SP hoped to raise that percentage to between 30 and 50 percent. According to his testimony, it was standard procedure that any time a clerk estimated a weight on the waybill, some notation on the waybill was needed to indicate that the weight was estimated. He further testified that more and more shippers were dealing directly with the billing office in Los Angeles rather than dealing with yard clerks in the various outlying areas.

Calnev Pipeline Dispatcher Training Program

The primary function of a Calnev pipeline dispatcher was to operate and monitor the pipeline through use of a computer-based operating system. This computer system monitored the condition of the pipeline and incorporated several safety mechanisms that would automatically shut down the system in the event of an emergency.

According to Calnev's manager of operations, there were no written criteria the company followed in selecting an individual for the position of dispatcher. The employee turnover rate was low, and individuals filling the positions of dispatcher normally came from within the company and were knowledgeable of Calnev's operations and procedures.

A trainee received an overview of the Calnev pipeline system and was then paired with the on-duty day shift dispatcher, who was responsible for the trainee's on-the-job training. The duration of on-the-job training varied with the individual. According to the manager of operations, an individual experienced in Calnev's operations might only require 2 months of on-the-job training before being allowed to dispatch while other individuals who were not as knowledgeable might require up to 6 months of on-the-job training.

The on-duty dispatcher provided updates on the trainee's performance to the terminal supervisor and the manager of operations. After a 6-month period, a trainee received a written performance appraisal. After a trainee had completed on-the-job training and had shown a competent working knowledge of the system, the dispatcher was monitored while operating the system alone. Performance was monitored continually by an event recording system, which recorded every keystroke entered on the computer by the dispatcher and all alarms received during the employee's shift. The event recorder printout was reviewed by company officers after an occurrence involving unusual circumstances.

To supplement on-the-job training, the trainee was exposed to several on-going training programs. These programs included monthly meetings concerning safety and operations, review and completion of the operator training manual, and special training seminars. The operator training manual was a self-paced, self-instructional two-volume document that covered a wide variety of pipeline operational procedures. Trainees reviewed these manuals while on duty, a chapter at a time. When the individuals believed they had adequately reviewed the chapter, they were examined on the material. A company officer administered the exam and reviewed all incorrect responses with the trainees. Trainees were to complete all chapters and associated tests during their first year of employment.

The dispatcher on duty at the time of the rupture received his 6-month performance appraisal on March 30, 1989, with the rating of "meets most performance requirements." His instructor had described the dispatcher's ability to learn material as "slow" at that time but attributed this to the dispatcher's refinery rather than pipeline background. The instructor added that as time passed, the dispatcher "quite easily" learned the proper operating and dispatching procedures.

Southern Pacific Management Oversight of Train Operations

The SP's road foreman of engines was responsible for the direct supervision of engineers operating over his particular territory. The road foreman of engines, whose territory was involved in the train derailment, testified that he was responsible for 35 to 55 engineers, depending on the number of helper units in service and the amount of train traffic. According to his testimony, in addition to the required rules examinations, rules compliance was measured through efficiency testing, train rides, review of event recorders, and general observation.

The road foreman of engines for the territory involved in the train derailment testified that efficiency tests were conducted 7 to 8 days a month and that 50 percent of that time would be devoted to checking speed violations through use of radar. The other 50 percent was devoted to efficiency testing of other operating rules. According to the road foreman, there was no set policy on the number of efficiency tests to be made on grade operations or through the use of radar. With respect to train rides, the road foreman testified that he would ride with each engineer at least once or twice a year or more if the engineer was experiencing problems. Again, there was no written policy regarding the number of check rides that had to be made. According to the road foreman, he reviewed 15 to 20 speed tapes a month, some of which were reviewed with the engineer if the road foreman had some concern about the engineer's performance.

The SP instituted a demerit system for rules violations as one method of disciplinary action. According to the road foreman, an employee could accumulate up to 90 demerits before suspension or disciplinary action was initiated. He stated further, however, that if an employee had accumulated 60 demerits, an assessment of the employee's performance was made. For each month that no violations were incurred, two and one-half demerits were removed from the employee's record.

SP's records indicated that in the 12 months prior to the train derailment, the head-end engineer had successfully passed 68 of 70 efficiency tests conducted. His records indicated two instances of disciplinary action. On March 31, 1986, he was cited for exceeding maximum authorized speed (29 mph in a 25-mph zone) while serving as fireman during helper engine service. He waived a formal investigation and received 30 demerits. The second instance involved his failure to properly connect locomotives on February 13, 1988. Again, he waived a formal investigation and received 30 demerits.

SP's records indicated that in the 12 months prior to the train derailment, the helper engineer had successfully passed all 63 efficiency tests conducted. His records indicated no instances of disciplinary action.

None of the crewmembers involved in the train derailment on May 12, 1989, were cited for disciplinary action. According to the general manager for the Western Region, one reason for not taking any disciplinary action was because of the false information provided to the traincrew. He testified, "...it would not have seemed appropriate due to all the outside factors to cite this crew....It would have been very difficult to establish the complicity of the crew as far as the runaway train."

Industry Pipeline Standards and Federal Regulations

When the construction of the Calnev pipeline began in 1969, there were no Federal regulations in effect that addressed the operation, inspection, and maintenance of liquid pipelines. Industry-recommended standards, American Standards Association (ASA) Code B31.4 - "Liquid Petroleum Transportation Piping System" (as revised in 1966), addressed design, construction, inspection, testing, operation, and maintenance considerations, which liquid petroleum operators were encouraged to follow. Selected provisions of the code are contained in Appendix I.

Federal authority to regulate liquid pipeline carriers for safety purposes has existed since March 4, 1921, and was vested originally in the Interstate Commerce Commission (ICC). In 1967, this authority was transferred to the FRA of the U.S. Department of Transportation (DOT), and shortly thereafter, the first Federal safety regulations for liquid pipelines were issued requiring only the reporting of accidents (49 CFR 180.28).

In August 1968, the Natural Gas Pipeline Safety Act of 1968 was enacted, and the Office of Pipeline Safety (OPS) within the DOT was established to develop safety standards for natural gas pipelines and to provide technical advice to the FRA on matters relating to liquid pipelines. On September 29, 1969, the FRA issued regulations for liquid pipelines, 49 CFR Part 195. (The regulations did not apply to pipelines already constructed or under construction.) Many of the provisions of the regulations were based on the existing industry standards, including the 1966 edition of the ASA Code B31.4. Pertinent provisions of Part 195 are contained in Appendix J. Only a few substantive changes have been made to these particular provisions since the regulations were issued in 1969.

ASA Code B31.8, "Gas Transmission and Distribution Piping Systems," is the industry standard for the natural gas industry. Code B31.8, unlike Code B31.4, had established design standards based on the surrounding population. In determining the population density, the number of buildings intended for human occupancy within a 1/4-mile exposure distance on each side of a gas pipeline route was to be considered. Initially, these standards applied only to the original installation of pipelines, and modifications were not required when the population adjacent to the pipeline increased. However, the 1968 edition of Code B31.8 recommended that gas pipeline operators continually survey their pipelines, and that for pipelines operating in excess of 40 percent of the specified yield strength of the pipe, operators confirm the adequacy of the design or reduce pressure in the pipeline when prescribed population densities were exceeded. Additionally, Code B31.8 (as revised in 1968) based the frequency of several tests required for acceptance of newly installed pipeline, and of several inspections required of pipelines in operation, on the population densities adjacent to a pipeline.

The first Federal regulations for natural gas pipelines, 49 CFR Part 192, were published on August 19, 1970, and were primarily based on the 1968 edition of Code B31.8. Pertinent provisions of Part 192, specifically the population-based spacing requirements for valves on natural gas transmission lines, are contained in Appendix K.

Oversight of Calnev's Pipeline Operations

The Calnev pipeline involved in the train derailment and the subsequent pipeline rupture is an interstate liquid pipeline. Federal regulations addressing interstate pipelines, as contained in 49 CFR Part 195, are currently administered by OPS within the Research and Special Programs Administration (RSPA), a part of the DOT.^[31] The Office of the California State Fire Marshal (CSFM) has authority for the regulation, inspection, and enforcement of intrastate pipelines. On January 1, 1987, the CSFM signed an agreement with OPS that stipulates that the CSFM will act as an agent for OPS for inspecting and monitoring interstate pipelines within the State of California to determine compliance with certain provisions of 49 CFR Part 195. Because construction of the Calnev pipeline began in 1969, the provisions of 49 CFR 195 were not in effect; thus, the design, materials, installation (including the location of valves), and initial testing requirements do not apply to this pipeline. However, the provisions for reporting accident and safety-related conditions and for the operation and maintenance of the pipeline do apply.

As an agent for OPS, when CSFM detects a violation of 49 CFR 195, it advises OPS of the findings. Based on its review of the information provided by CSFM, OPS determines if enforcement action is warranted, the type of action warranted, and whether or not to pursue further action. According to a representative from the CSFM, in this arrangement, CSFM serves to detect noncompliance but has no regulatory authority in resolving any noncompliance detected. Testimony from the division chief for pipeline safety operations at CSFM indicated, however, that CSFM could request an operator to take corrective action without first consulting OPS if an immediate risk to public safety existed.

The San Bernardino deputy fire chief (incident commander) testified that although he had been contacted by a representative from the CSFM on the day of the derailment, he was not made aware of the presence or activities of the CSFM during the days following the train derailment. Testimony from the division chief of pipeline safety operations indicated that representatives from the CSFM were on site through May 16, were in contact with Calnev personnel throughout this time concerning cleanup operations and inspection of the pipeline, and relayed information concerning activities at the derailment site to the OPS' regional office in Colorado. According to his testimony, OPS did not instruct CSFM to take any actions at the site, CSFM representatives on site were satisfied with Calnev's inspections, and based on Calnev's assessment of the integrity of the pipeline, CSFM did not request Calnev to take any further action. He stated also that CSFM was not aware of any request by the deputy fire chief to fully expose and inspect the pipeline in the derailment area. The division chief further testified that representatives from CSFM routinely worked with pipeline personnel rather than fire department personnel, but that CSFM had initiated a program subsequent to the pipeline rupture to contact the fire departments within the State of California to inform them of CSFM's role in and responsibilities for liquid pipelines.

Following the pipeline rupture, representatives from the CSFM and from OPS were dispatched to the scene of the accident. The deputy fire chief stated that he was made aware of their presence and was routinely updated on their activities during the days following the rupture. (The actions taken by the OPS following the pipeline rupture have been previously discussed.)

On August 9, 1989, as a result of its preliminary investigation of the pipeline rupture, the Safety Board issued the following two Safety Recommendations to the Research and Special Programs Administration:

P-89-5

Require pipeline operators that have "All-Clear Check Valves" manufactured by the Wheatley Company installed in their pipeline systems to test these valves for proper closure and require the replacement of any that fail to close properly.

P-89-6

Establish inspection, maintenance, and test requirements to demonstrate and maintain the proper functioning of check valves installed in pipeline sytems.

On November 13, 1989, RSPA responded to the Safety Board's recommendations stating:

An Alert Bulletin has been issued that alerts all hazardous liquid pipeline operators to test in critical locations all check valves for proper closure and recommends the replacement of any check valve that fails to close properly. Also, the advisory recommends that valves located in noncritical areas be inspected for operation at the first opportunity the valves can be bypassed or otherwise taken out of operational service. (The full text of the alert bulletin is contained in appendix L.)

We have initiated a study to determine the feasibility of establishing inspection, maintenance, and test requirements to demonstrate and maintain the proper functioning of check valves installed in pipeline systems. We plan to complete this study within 9 months. If the study supports a need for such a regulation, we will initiate rulemaking.

Based on RSPA's response to the Board's recommendations, Safety Recommendations P-89-5 and -6 have been classified as "Open—Acceptable Alternate Action" and "Open—Acceptable Action," respectively.

Meteorological Information

At 7:30 a.m. on May 12, 1989, at the Norton Air Force Base, located about 4 miles from the accident site, the sky was clear with a temperature of 57 degrees F. Visibility was reported as 15 miles. Similar weather conditions existed at the time of the accident site.

Medical and Pathological Information

Train Derailment

Two children, ages 7 and 9, suffered fatal injuries when the train derailed and hopper cars struck their house at 2348 Duffy Street (see figure 11). Postmortem examinations indicated that both children died of suffocation and compressional asphyxia.

The head-end engineer of Extra 7551 East sustained a 4-inch laceration of the left upper arm, multiple rib fractures on the left side with pneumothorax, and multiple abrasions and contusions. He was admitted to the intensive care unit at St. Bernardine Hospital where he was treated and later released.

The two crewmembers located in the last helper engine reported receiving minor injuries. Immediate medical attention was not sought, and there are no records to indicate injuries or treatment.

A resident at 2326 Duffy Street (see figure 11) sustained multiple injuries, including a right compound fracture of the femur, a large laceration of the right knee, and a compressed spinal fracture when several hopper cars struck his house. This resident was trapped for about 15 hours before being rescued and transported to a local hospital.

The conductor of Extra 7551 East, who was located in the lead engine unit, 8278, and the brakeman who was located in the third engine unit, 7549, suffered fatal injuries as a result of the derailment. Postmortem examinations indicated that both crewmembers died of multiple traumatic injuries.

Pipeline Rupture

Two residents, one of whom was in her house at 2327 Duffy Street and the other in her backyard at 2315 Duffy Street (see figure 11), sustained fatal injuries as a result of the fire.

Three residents received serious injuries, second and third degree burns, while escaping from their burning homes. Sixteen other residents reported minor burns and shortness of breath from smoke inhalation. One firefighter reported burning his foot while fighting the fire.

One person, who was not a local resident, received multiple rib fractures in an automobile accident while attempting to make a U-turn to avoid the fire resulting from the pipeline rupture. Three other persons, who also were not local residents, reported minor injuries, including lacerations and contusions, while attempting to drive away from the fire.

Toxicological Information

In accordance with current FRA requirements, toxicological samples were obtained from all five crewmembers of Extra 7551 East. These samples (blood and urine specimens from the surviving crewmembers,^[32] and blood, urine, and tissue specimens from the deceased crewmembers) were forwarded to and examined by the Center for Human Toxicology (CHT) in Salt Lake City, Utah. Additionally, in accordance with SP operating procedures, a second urine specimen was collected from each of the surviving crewmembers and forwarded to an alternate contract laboratory facility, Roche Biomedical Laboratories, Incorporated (RBL), for examination. The specimens examined by CHT and RBL were negative for alcohol and other drugs.

The train dispatcher on duty at the time of the train derailment was not requested to submit to toxicological testing. Calnev's pipeline dispatcher on duty at the time of the pipeline rupture was not requested to submit to toxicological testing. Calnev did not have a policy regarding postaccident toxicological testing of employees. Calnev employees, however, were required to submit to drug testing before being hired. Testimony by Calnev's manager of operations indicated that Calnev was aware that the company would be required by Federal regulation to implement a drug testing program in the near future.

Southern Pacific's Physical Examination Policy

SP's physical examination policy requires all employees to submit to a physical examination when they are hired. With the exception of engineers, there is no requirement that employees submit to further examinations after that date. Engineers must undergo a physical examination at the time they are promoted to the position of engineer. They are not required to submit to another examination until they reach the age of 40, at which time they must then undergo a physical examination every 5 years until the age of 60. At 60, an engineer must then receive an annual physical examination. At age 65, engineers are required to undergo semiannual examinations.

Tests and Research

Event Recorders

The multi-event recorders recovered from head-end locomotive units 7549, 7551, 8278 were sent to the Safty(sic.) Board's laboratory in Washington D.C., for readout and evaluation.

The type of recorders installed on the SP locomotive units involved in the accident were designed to record speeds up to 90 miles per hour (mph). The three stripcharts generated from the event recorders indicated that the train speed exceeded 90 mph. Because the physical limit of these stripcharts was exceeded, the maximum speed of the train could not be determined based on the original recorded values. To determine the maximum speed attained, additional stripcharts were generated using a method that reduces the recorded speed values to half their original values (appendix M). Actual values at any point on the stripchart are then obtained by doubling the indicated speed.^[33] The results indicate that the train probably reached a speed of 110 mph before derailing.

By reviewing the stripchart generated from the information recorded from unit 7549, Safety Board investigators attempted to determine if the dynamic braking on that unit was functioning. If the dynamic brakes on a locomotive unit are functioning, whenever an engineer uses dynamic braking, corresponding amperage activity should occur and be recorded on the stripchart. A review of the stripchart indicated that unit 7549 went into dynamic braking on 15 occasions during the previous 30 hours of operation; however, the expected corresponding amperage activity was recorded on only 2 occasions. Both instances of recorded amperage activity occurred before Extra 7551 East reached Hiland. The SP chief mechanical officer testified, "...I do not have the degree of confidence in the reconstructed tape that [the general road foreman] does because of the difficulty we've experienced with the tape cartridges. It's not uncommon to have them not record on a channel." The general road foreman testified that based on his review of the stripchart for unit 7549, "During the time that the train descended the hill from Highland, the dynamic brake did not work."

The event recorder printout indicated that service braking (air/mechanical brakes) occurred for more than 25 minutes as the train descended the hill from Hiland. According to information obtained from a brake shoe manufacturer, "Composition brake shoe binders start to decompose at temperatures between 700 degrees F and 800 degrees F, provided this elevated temperature is sustained. If composition brake shoe temperatures are sustained for an extended period of time (20 minutes or greater) above 700 degrees F and decomposition takes place, the shoe will continue to produce high frictional values with small losses as the result of heat fade."

Train Dynamics Analyzer Runs

On August 15, 1989, six simulations of the movement of train Extra 7551 East down the 2.2 percent grade from Hiland were conducted on a Freightmaster Train Dynamics Analyzer in Fort Worth, Texas. Operating parameters, including air brake reductions and speeds, were based on the information contained on the stripchart made from the event recorder data pack removed from SP 7551 following the derailment. As stated by SP's general road foreman, who observed the simulations with Safety Board investigators, "Test one is the only test that we could run that would allow us to go down the hill in the same manner that this train went down the hill and make the air brake reductions as they were made on the strip chart." Test one was made with 12 axles of dynamic braking on the head-end locomotive units, 6 axles of dynamic braking on the helper units, and with a trailing tonnage of about 8,900 tons. The brake shoe efficiency was purposely degraded during the run with the level of degradation and the location of degradation estimated as follows:

Mile Post Location	Percent Brake Shoe Efficiency
469	75
473.7	60
474.7	55
475	50
480.7	40

The general road foreman of engines recounted the results of the simulation, "We maintained the 30 miles an hour with the reductions that was(sic) made on the strip chart and then as the speed started increasing on the strip chart, we started brake deterioration in the simulations and things deteriorated from that point on...the train obtained approximately 105 miles per hour."

Test four was conducted with 12 axles of dynamic braking on the head-end of the train, 12 axles of dynamic braking on the rear end, and a trailing tonnage of about 6,150 tons. These parameters represent the number of axles of dynamic braking and the tonnage that the head-end engineer believed existed for Extra 7551 East. The simulation revealed that the train was controlled and the speed maintained under 30 mph coming down the hill.

The other four tests were stopped when the train could not be controlled coming down the hill by using the parameters from the event recorders.

Instrumented Brake Shoe Tests

On June 12, 13, and 14, 1989, SP conducted brake shoe tests on SP cars equipped with empty/load devices and on DRGW cars not equipped with the devices. The tests were conducted to determine braking forces on cars similar to the cars that were in the accident. By replacing the actual brake shoe with an instrumented brake shoe, accurate measurements of the forces applied to the wheel could be made. According to the SP's chief mechanical officer, the tests confirmed that the SP cars had "...a braking ratio of 1...."

Train Vibration Study

At the request of the Safety Board, the Test and Engineering Center of Failure Analysis Associates, an engineering and scientific consulting service, conducted tests at the accident site to measure and record vibration and strain levels to determine if the passage of trains induced vibration or strain in the buried pipeline. As stated in the introduction to the report prepared by Failure Analysis Associates, "...an instrumentation system was assembled to provide a measure of the vertical and lateral acceleration at two locations and axial and hoop strains at two locations on the pipe. Data were acquired for a 24-hour period during which time nine trains passed through the area. In addition, consist and engine log data were acquired from the SP for several of these trains. After analyzing the data collected, Failure Analysis Associates concluded, "...it does not appear that the passage of trains, at the speeds observed, imparts a measurable strain or vibration in the pipeline."

Soil Inspection Report

On May 25, the day of a pipeline rupture, Calnev contracted with Converse Consultants, a geo-technical and environmental consulting organization, to perform work in the area of the pipeline rupture. As stated in its August 30, 1989, report of findings (appendix N), Converse Consultants' investigation "...was performed to evaluate the subsurface conditions in the vicinity of the pipeline rupture in order to locate areas where the soils may have been disturbed by excavating equipment. It is our understanding that excavating equipment may have been utilized in the vicinity of the pipe rupture during Calnev post derailment pipe inspection and/or during clean-up of the derailment debris." A total of 14 tests were conducted; tests 1 through 4 (figure 16) were performed within the area of the rupture, and tests 5 through 14 were conducted in an area ("control area") where Converse Consultants believed there had been no excavation or disturbance of the soil. According to Converse Consultants' report, tests of samples taken at locations 1 through 4 indicated "...disturbed or poorly compacted earth materials...and contained significant quantities of the mineral trona." Tests of samples taken at locations 5 through 14 indicated that the earth materials had not been recently disturbed. The tests indicated no presence of the material trona at these locations. A representative from Converse Consultants testified, "...my interpretation and conclusion is that the materials, backfill materials, which prior to the derailment would have been just clean, natural soils without the presence of trona, had become contaminated with trona by means of excavation and replacement, probably as backfill or certainly as materials that had been exposed to trona and mixed, by whatever means."

Metallurgical Testing

Two 14-inch outside diameter (OD) pipe sections, one measuring 44 inches long and containing a rupture and one about 41 inches long, were taken to the Safety Board's materials laboratory in Washington, D.C., for examination. The two sections of pipe had been adjacent to each other before they were cut apart. As received in the Board's laboratory, the pipe contained directional arrows and a marking along the top of the section to indicate the orientation of the pipe in the ground before removal. Arrows "N" and "S" denoted the north and south directions, respectively (figure 17). A longitudinal marking across the sections at the transverse cut signified the top of the pipe and the matching rotational positions of the two sections relative to each other. Yellow grid line markings had been made on the OD surface around the rupture area. Subsequent notes supplied by Failure Analysis Associates (the metallurgical consultants contracted by the SP to examine the pipe) indicated that the markings denoted positions where thickness measurements had been made on the pipe. Arrow "x" in figure 18 indicates a location where the wall thickness measured the thinnest at about 0.249 inch, which was confirmed by micrometer measurements in the Safety Board's laboratory. Wall thicknesses of 0.254 inch were also found in the origin area of the fracture. The wall thickness away from the fracture measured about 0.312 inches.

The northern section of pipe contained a gaping rupture on the east side of the pipe (bracket "o" in figure 18). As shown in figure 19, the fracture faces were gaped apart and the pipe was deformed outward.

Examination of the OD surface of the pipe sections disclosed what appeared to be mechanical damage in the form of depressions or scrapes which, for the most part, were linear. The most severe damage was on the northern section of pipe and in line with the origin of the rupture. Unmarked arrows in figure 18 outline the damage, which was readily visible. This damage produced a visible depression in the pipe OD surface with a matching bulge on the inside diameter (ID) surface. The maximum depth of the depression was estimated to be about 0.18 inch from the original OD shape. The width of the damage was about 2 inches at its maximum point.

Arrows in figure 20 outline mechanical damage to the OD surface on the southern section of pipe. This section contained two pronounced areas of elongated damage, the centers of which were 2 to 3 inches apart. Neither of these areas showed appreciable denting into the OD surface.

Visual examination of the fracture surface of the rupture disclosed no evidence of progressive cracking. All fracture features were typical of an overstress separation. A pie-shaped section containing the origin area of the rupture was excised from the pipe and further sectioned to a specimen size suitable for examination with the aid of a scanning electron microscope (SEM). SEM examination disclosed dimple rupture features throughout the fracture area that were typical of a ductile overstress separation. There was no evidence of crack arrest markings or oxidation areas that would indicate a progressive separation.

Many parallel microfissures were noted on the outside diameter in the origin area near the fracture plane. Most of these microfissures were extremely small and shallow and, for the most part, detectable only by higher magnification. However, some microfissures were readily visible with the unaided eye. SEM examination of the fractures within these larger microfissures disclosed features also representative of an overstress separation.

To better characterize the mechanical damage to the OD surface, several metallographic sections were prepared that were oriented both transversely and in line (along the length) with the linear depression. Arrows "B" and "C" in figure 18 indicate the general area where these sections were prepared. The sections were etched and examined along the OD surface for evidence of grain distortion. Except for sporadic highly isolated areas, there was no evidence of grain distortion that would signify a direction of deformation. A few very small areas were noted along the OD surface that were indicative of particles impacting the OD surface radially inward with a slight sliding movement. There was no evidence of grain distortion that would indicate a massive movement of the material in the depression.

A section of pipe located south of the rupture and which contained two areas of surface damage—one near the top centerline and one on the west side—was sent to the Southwest Research Institute for metallurgical examination. The principal objectives of the examination were to inspect for the presence of cracks and to identify the direction of surface deformation in the two damaged areas on the sample. A summary of the results follows:

1. No evidence of any surface cracking was observed on the outside surface of either sample.
2. No significant wall thinning had occurred in either of the scraped areas. The minimum wall thickness measured at the point of most severe damage 0.313 inch, while the undamaged wall thickness was 0.317 inch.
3. The pipe had been locally dented inward approximately 0.1 inch at the damaged area near the top centerline (southernmost damage area).
4. SEM and EDS analyses of the surfaces did not detect any tool-to-pipe metal transfer.
5. Metallographic sectioning positively identified the direction of surface deformation in both areas of damage.

   a. Damage near top centerline

   The direction of surface deformation was established to be in a mainly southerly direction.

   b. Damage near 270 degree position (west side)

   The direction of damage was established to be in a downward and southerly direction. This direction is consistent with the nature of the coating damage.

Simulation of Excavating Equipment Operations

On January 6, 1990, in accordance with a test plan agreed to by all parties, Calnev conducted a series of field tests to determine the amount of damage that three pieces of excavating equipment could inflict on a 14-inch pipeline. These three pieces of equipment that worked in the vicinity of the pipeline between May 12 and 19, 1989, following the removal of the train cars and locomotives, were a Case 580C rubber tire backhoe, a John Deere 690 track excavator, and a Caterpillar 988B front-end loader.

Two 80-foot lengths of pipeline that had to be removed from the accident site were filled with water and pressurized to 800 psig and buried without anchors to about minimal burial conditions (one was buried to a 4-foot depth and the other to a 1 1/2-foot depth) that might have been encountered in the area of the train derailment during cleanup operations. The backhoe and the excavator were owned and operated by the Arizona Pipeline Company, and the front-end loader was owned and operated by Jimco Construction Equipment Company, working on behalf of SP. In addition to Safety Board personnel, representatives from Calnev, the Southern Pacific Transportation Company, the California State Fire Marshal's office, IT Corporation, and the Office of Pipeline Safety were present for these field tests.

The teeth on the 2-foot-wide bucket of the Case 580C backhoe penetrated the pipeline coating but could not substantially dent the pipe wall in any of the tests. Running the teeth of the bucket along the top of the pipeline resulted in shallow "chatter" type scratches in the pipe wall. The bucket of the backhoe, with teeth down, was pulled across the top of the pipeline at various angles; pulling the bucket across at an angle of 45 degrees resulted in the greatest penetration to the pipeline coating and the pipe wall with all five teeth of the bucket. Dropping the bucket from a 6-foot height and a 2-foot height and hitting the pipeline with the back of the bucket did not result in any dents to the pipe wall. Because the hydraulics of the equipment slowed the bucket speed when dropped from the 6-foot height, the damage to the coating was less than the damage that occurred when the bucket was dropped from the 2-foot height. The teeth of the bucket did not penetrate or dent the pipe wall when dropped onto the pipeline.

Running the teeth on the bucket of the John Deere 690B excavator along the top of the pipeline resulted in chatter type marks in the pipe wall similar to those made by the Case 580C backhoe. Scraping the side of the pipeline with the side of the bucket resulted in damage to the pipeline coating but no dents in the pipe. Two hits on the pipeline with the back of a loaded bucket created a dent about 1/16-inch deep in the top of the pipe.

During the first test on the second piece of buried pipeline using the Caterpillar 988B front-end loader, the operator dug into the soil covering the pipeline and then dragged the back of the bucket over the top of the pipeline. The operator stated that he did not feel the equipment hit the pipeline, and there was no noise at ground level of the equipment striking the pipeline. After the pipeline was uncovered by hand at this location, observers saw that two marks physically disturbed the metal, about 2 feet apart, on the top of the pipeline. Also, coating damage was observed. A second attempt to drag the back of the bucket over the top of the pipeline resulted in distinctive marks, 18 inches apart, to the coating and the pipe wall. During this second attempt, the operator felt the equipment hit the pipeline, and the noise of the equipment striking the pipeline was clearly heard at ground level. When the side of the bucket was forcefully scraped along the side of the pipeline in a forward motion, damage to the pipe coating was extensive. Where the coating damage ended, a tooth of the bucket struck the lower quadrant of the pipeline creating a deep dent. This action also caused the unanchored pipeline to move 4 inches in a longitudinal direction. When the side of the bucket was scraped along the side of the pipeline a second time over a 5-foot length of the pipeline, a 4-inch-wide area of coating was removed along the entire length. When the back of the bucket of the front-end loader was dragged over the top of the pipeline a third time, two marks, 5 inches apart, were observed along the top quadrant of the pipeline. There was no visible denting of the pipe at these locations.

Other Information

Train Movements Following the Train Derailment and Preceding the Pipeline Rupture

Between the time the SP opened its rail line for traffic at 4:00 p.m. on May 16, 1989, and the time of the pipeline rupture on May 25, 1989, 34 trains and 1 light engine were operated eastbound, and 39 trains and 1 light engine were operated westbound.

Agreement Between the Southern Pacific and City of San Bernardino Following the Train Derailment

An agreement between the Southern Pacific and the City of San Bernardino relative to the train derailment of May 12, 1989, was presented at the Safety Board's public hearing in August 1989 (appendix O). In addition to outlining the obligations of the railroad with respect to the property destroyed or damaged as a result of the train derailment, the agreement provided that Southern Pacific, rather than the City, would be responsible for any reimbursement claims by Calnev. The agreement further stated:

It is further hereby acknowledged and agreed by the parties that a Cal-Neva^[34] gas line runs adjacent to the location of the derailment; that the health, safety and welfare of the persons in the vicinity of the derailment requires that such line be fully exposed to allow visual and other examination to the satisfaction of the City Fire Department. As between City and Railroad, Railroad shall bear all costs incurred thereby and for replacement of the line. Railroad's obligation to Cal-Neva shall be determined by the contract between Cal-Neva and Railroad, if any.

This agreement may be amended only in writing by and between the parties hereto.

The agreement was signed on May 17, 1989, by the general manager of SP's Western Region and the City Attorney for San Bernardino.

The deputy fire chief (incident commander), who stated that he had expressed the desire to Calnev's manager of operations during the immediate days following the train derailment that the pipeline be fully exposed and inspected, testified that he was not made aware of the provision of the agreement until June 21, 1989. According to his testimony, it was his understanding that he did not have the authority to require Calnev to expose and inspect the pipeline and that only the State Fire Marshal's Office through the Office of Pipeline Safety had that authority. The deputy fire chief stated that he did not make his desire known to the State Fire Marshal's Office. The deputy fire chief terminated his command of the emergency response to the train derailment on May 15, 1989.

The general manager of SP's Western Region testified that when he signed the agreement, it was his belief that the inspection outlined in the agreement had been performed. Calnev's manager of operations testified that he was not aware of any agreement between the City and SP regarding the exposure and inspection of the pipeline and that there had been no contract between Calnev and SP. He testified also that, based on his understanding of the right-of-way agreement between SP and Calnev, SP could have requested Calnev to expose and inspect the pipeline. Testimony from the SP's general manager indicates that a request to fully expose and inspect the pipeline was never made to Calnev.

Development of Land Adjacent to the SP Railroad and the Calnev Pipeline

The area affected by the May 12 derailment and the May 25 pipeline rupture was planned in 1955 for residential use, and the subdivision plat was recorded with San Bernardino County on November 10, 1955. On October 1, 1957, the subdivision was annexed by the City of San Bernardino and incorporated within the city limits. In 1967, the SP constructed the portion of its railroad where the train derailment occurred, and at that time, no houses were located on Duffy Street.

By October 1967, houses had been constructed within the eastern portion of the subdivision, but no houses were on either side of that portion of Duffy Street that paralleled the proposed railroad. In 1969 and 1970, when the Calnev pipeline was constructed along the eastern edge of the SP right-of-way, no houses had yet been erected on that portion of Duffy Street that paralleled the railroad; only a few houses had been built within the subdivision. According to recollections of long-term residents, intensive construction within the area occurred from 1970 to 1970.

The City of San Bernardino's General Plan for land use is a policy document that establishes goals, objectives, and policies for the future. The specific standards for a development are to be guided by this Plan and included in the zoning ordinances or development codes. The subject of land use control because of its proximity to railroad mainline tracks or to high pressure liquid or other pipelines is not specifically addressed.

Before these accidents, the City had developed a proposed revision to its Plan, subsequently conducted public hearings on the proposal, and approved a revised plan. A statement within the proposal advised that, in part, this plan is a foundation policy document that defines the framework for decisions by the City on the use of its land for the protection of residents from natural and human-caused hazards. Neither the proposal nor the newly adopted plan specifically addressed the use of land near mainline railroads or high pressure pipelines.

Disaster Preparedness

San Bernardino County, about 20,000 square miles in size, is located in the southeastern portion of California. Within the county are 20 incorporated cities with the heaviest concentration of population in the west-central portion. The county's population is more than 1 million.

The County of San Bernardino, the district fire agencies, and the municipal fire departments are signatories to the State of California's Master Mutual Aid Plan to combat emergency situations that may develop and that are beyond the control of any one agency. In addition, many of the agencies have developed local mutual aid and automatic aid agreements. To maximize the resources within the County and to assist in the coordination of such resources, a Mutual Aid System was developed that divides the County into 10 zones. The SP train derailment occurred in what is designated in the Mutual Aid Plan as Zone 2.

Zone 2, or the "East Valley" area is served by eight agencies in the east end of the San Bernardino Valley (figure 21). Resources of the agencies in Zone 2 include: 83 fire response vehicles, 28 specialty units and squads, and 6 pieces of specialized equipment. Within Zone 2 are 526 full-time firefighters and 25 reserve firefighters.

The San Bernardino County Communications Center located in Rialto servers as the Zone 2 Emergency Communications Center. The Communications Center is responsible for emergency dispatch functions for the San Bernardino County Fire Agency-Central Valley District and the Rialto and Loma Linda Fire Departments. Separate dispatch centers are maintained by the fire departments of the City of San Bernardino and Norton Air Force Base, and by the County Fire Warden.

Train Derailments over Pipelines

The California State Fire Marshal's Office has maintained records on pipeline failures since it began regulating hazardous liquid pipelines in 1984. On March 9, 1989, a butane car derailed at the Tosco Refinery in Martinez, California, and struck and ruptured an above-ground pipeline. No injuries, fire, or explosion resulted from the accident. In another recorded incident at Montclair, California, on December 19, 1988, an axle from a "rail car truck" had made a small hole in the 20-inch-diameter pipeline of the Southern Pacific Pipe Line Company; the pipeline ran parallel to the railroad tracks.

On June 27, 1989, a locomotive was being used to switch the order of rail cars at a Union Pacific Railroad yard at Las Vegas, Nevada. About 8:30 a.m., Pacific daylight time, 34 rail cars were being moved when the leading 9 cars and the trailing 12 cars derailed with several rail cars overturning on top of two Calnev petroleum products pipelines. The 6-inch pipeline located on one side of the rail line contained jet fuel, and the 8-inch pipeline on the opposite site of the rail line contained gasoline. Both pipelines were under about 600 psig pressure and both were buried 4 to 5 feet below the ground surface.

Pipeline inspection personnel from both the Nevada Public Service Commission and the Office of Pipeline Safety responded to the Las Vegas accident to monitor the removal of rail cars, to require inspection of both pipelines to determine if the pipelines had been damaged, and to determine if they were safe to return to service. The Office of Pipeline Safety required Calnev to fully uncover and visually inspect the pipelines for possible damage and then required Calnev to hydrostatically test the pipelines through the area of the derailment. The Office of Pipeline Safety advised the Safety Board that it had established as a policy that pipelines potentially damaged by a derailment would be both visually examined and subjected to a hydrostatic test before they could be returned to service, if OPS believes there is potential for harm to life or property.

The Safety Board requested that the Santa Fe Pacific Pipelines Company (formerly the Southern Pacific Pipelines Company)^[35] provide records of any derailments over pipelines and their results. Santa Fe advised that 55 percent of its 3,300-mile pipeline system was installed along railroad rights-of-way and that between 1966 and 1989, 121 train derailments had occurred over its pipeline. The Santa Fe has never experienced any damage as a result of a train derailment where the pipe was buried 3 feet or more below ground. However, it did experience damage to its pipeline during the derailment clearing operations for the Montclair accident.

On June 20, 1989, the California Senate Committee on Toxics and Public Safety Management and the California Assembly Select Committee on Hazardous Materials and Pipeline Safety held a joint public hearing on the San Bernardino accidents. As a result of that hearing, Assembly Bill No. 385 was passed and signed into law. The bill calls for the California State Fire Marshal to conduct and prepare a risk assessment study addressing hazardous liquid pipelines within 500 feet of a railroad track. The study is to be completed by January 1, 1991.

Notes

1. A raw material composed of sodium carbonate, sodium bicarbonate, and water. It is a source for soda ash, pure sodium carbonate, and is used in the manufacture of fertilizer.
2. The purpose in "releasing" or changing the status of a car is to release the customer (in this case Lake Minerals) from the per diem charge for holding empty cars.
3. Southern Pacific’s computer system is composed of various files including a car file and a waybill file. Additional discussion occurs under Method of Operation.
4. A document provided to the traincrew that indicates, among other information, the tonnage of the train.
5. In a unit train, all the cars are carrying the same product; for example, a unit coal train.
6. The route through the mountains over which SP trains often operate.
7. The SP air brake rules require that the train air brakes be tested before the train departs its initial terminal.
8. Train designation is based on the number of the lead locomotive unit. Even though unit SP 8278 was the lead unit in the locomotive consist, the train designation remained Extra 7551 East.
9. Dynamic braking is an electrical means used to convert some of the energy of a moving locomotive into an effective retarding force.
10. The method for positively determining if dynamic brakes are operating is by observing the amperage reading in each locomotive unit. See Mechanical Information for additional discussion.
11. At this location, the railroad tracks are constructed atop a 20- to 21-foot high embankment (levee).
12. A 14-inch liquid petroleum pipeline, operated by Calnev Pipe Line Company, was buried in the SP’s right-of-way.
13. A contract company (rather than a pipeline operating company, such as Calnev) that specializes in the installation, maintenance, and repair of underground lines.
14. Milepost numbers for the pipeline do not correlate with the milepost numbers for the railroad.
15. The static pressure in the pipeline varies with the elevation of the line. Therefore, the pressure reduction, rather than the pressure reading, was the critical observation at the two locations.
16. According to the Arizona Pipe Line Company employee operating the backhoe, all potholes were dug manually using shovels. According to Calnev’s maintenance superintendent, "The primary function of a pothole is to determine the depth and location of the pipeline. An excavation would be a larger hole, a more complete excavation where you are actually attempting to visually ascertain the condition of the pipeline."
17. A process by which vacuum-type excavation equipment makes about a 1-foot-diameter hole from ground level to the top of the pipeline.
18. The monitoring system at the Colton Terminal scans various pipeline parameters, including pipeline pressure, at 13-second intervals. Thus, an event (in this case, a pressure reading) may have occurred up to 13 seconds earlier than the recorded time (and the time cited in this discussion).
19. According to Calnev and OPS officials, the word "alarm" in the pipeline industry is not used to denote an emergency situation, but rather a change in operating conditions.
20. The dispatcher acknowledges the alarm by pressing a key on his computer terminal keyboard.
21. A small capacity pump activated first to bring the pressure up slowly to prevent surging when the mainline pumps are activated.
22. These persons were involved in a traffic accident while attempting to avoid the fire caused by the pipeline explosion.
23. With extended range dynamic brakes, as compared to standard range, more retarding force is available from 6 mph up to a speed between 18 and 25 mph depending on the gear ratio.
24. SP’s computer system contains a listing of all locomotive units and the status of any defects reported.
25. When an emergency air brake application is made, the PC switch, an electropneumatic device (sometimes referred to in the industry as the power cut-off switch or the pneumatic control switch), trips the eletric current which causes the main generators to unload and the engines to return to idle. When the air brace handle is placed in the handle off position, the PC will automatically reset. After the pressure is restored ithin 20 to 30 seconds (which the engineer can observe in front of him), the engineer can then manually move the handle and go back into dynamic braking.
26. As defined in the Air Brake Association’s Management of Train Operation and Train Handling, a retaining valve is "a control device through which brake cylinder air is exhausted completely or a predetermined brake cylinder pressure is retained." In short, the retainers provide the engineer with braking capability while the air brake system is being recharged.
27. Typically, when a company rule (in this case an airbrake rule) indicates that the superintendent will prescribe certain operating parameters, the superintendent will accomplish this through instructions in the timetable or by special bulletins.
28. Tonnage (or tons) per operative brake is computed by dividing the gross trailing tons by the number of cars in the train. The weight of the locomotives is not included in the gross trailing tons.
29. SP cancelled this rule by special instructions, effective May 22, 1989.
30. 6,151 tons divided by 88 (38 SP cars equipped with E/L devices figured at 1½ braking capability equals 57 (38 multiplied by 1½) plus 31 DRGW cars not equipped with E/L devices) equals 69 tons per operative brake.
31. On August 22, 1972, the U.S. Department of Transportation Act was amended to transfer the authority of the FRA to carry out the liquid pipeline safety functions to the Secretary of Transportation.
32. Samples from the head-end engineer, the helper engineer, and the helper brakeman were collected, respectively, at 12:15 p.m., 9:45 a.m., and 10:18 a.m., on the day of the accident.
33. Since the effect of the half-speed process on the other parameters is unknown, the stripcharts should be used to determine train speed only.
34. The Safety Board verified at the public hearing that the term Cal-Neva used in this agreement does refer to the Calnev Pipeline Company.
35. As a result of mergers subsequent to the Montclair, California, accident, Southern Pacific Pipelines became the Santa Fe Pacific Pipelines.