Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
The transportation of energy liquids and gases is often described as a high- hazard activity, characterized by low-frequency, high-consequence events. An especially tragic example is the July 2013 Lac-MÃ©gantic disaster, in which an unattended, runaway freight train crashed in the center of this rural town in eastern Quebec. Forty-seven people were killed, and 30 build- ings were destroyed by explosions and fires fed by more than 1.5 million gallons (~36,000 barrels) of Bakken crude oil carried in the unit trainâs 63 derailed tank cars.1 Pipeline and barge incidents have also had severe consequences. A July 2010 rupture of a corroded segment of transmission pipeline caused the release of nearly 850,000 gallons (~20,000 barrels) of viscous Canadian crude oil into a tributary of the Kalamazoo River in Marshall, Michigan, culminating in one of the largest and costliest inland oil spills in U.S. history.2 In March 2014, a bulk vessel collided with a 300-foot tank barge in the Houston Ship Channel causing a discharge of some 170,000 gallons (4,250 barrels) of fuel oil that sullied about 13 miles 1 Transportation Safety Board of Canada, âLac-MÃ©gantic Runaway Train and Derailment Investigation Summary,â 2014, http://www.tsb.gc.ca/eng/rapports-reports/rail/2013/r13d0054/ r13d0054-r-es.pdf. 2 National Transportation Safety Board, âEnbridge Incorporated Hazardous Liquid Pipeline Rupture and Release, Marshall, Michigan, July 25, 2010.â Accident Report NTSB/PAR- 12/01 (Washington, D.C., July 10, 2012), https://ntsb.gov/investigations/AccidentReports/ Pages/PAR1201.aspx. 4 Safety Performance of Long-Distance Crude Oil, Natural Gas, and Ethanol Transportation 68
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 69 of shoreline, endangered wildlife in several environmentally sensitive areas, and closed one of the countryâs busiest channels for several days.3 Preventing such tragedies is a major concern of industry, government regulators, and public safety officials; but so too is the prevention of smaller, less severe incidents because they can also be disruptive and costly to carriers, shippers, and communities. Smaller hazardous materials inci- dents may trigger a costly emergency response and be an indicator of safety risks not being properly managed. The long-haul bulk modes of freight transportationârail, pipeline, and waterâmove large amounts of hazard- ous materials on a daily basis, mostly without incident. All have done so for decades, long before the post-2005 growth in domestic crude oil, ethanol, and natural gas production. Accordingly, industry and government have a history of coopera tion in developing and improving systems intended to prevent hazardous materials incidents and to mitigate their consequences when they do occur. These systems consist of standards and procedures for the handling, labeling, and packaging of shipments; the design, fabrica- tion, repair, and testing of containers; the operations and maintenance of equipment and infra structure; and the identification and communication of shipment hazards and emergency response procedures. In addition, both industry and regulators monitor and investigate incidents, sponsor research and testing, and have outreach initiatives to inform carriers, shippers, and local authorities about safety standards and what to do in the event of a hazardous materials incident. Many large shippers and carriers maintain or hire specially trained emergency response teams and support industry-based mutual-aid networks that can assist in the response to hazardous materials incidents. It is against this backdrop of a broad-based system for assuring the safety of hazardous cargoes that the safety implications and risks of transporting larger volumes of domestic crude, ethanol, and natural gas movements must be considered. In the case of rail, the robustness of the safety assurance sys- tem has been tested by the rapid growth in crude oil and ethanol shipments, which until 10 years ago had a negligible rail traffic footprint. By compari- son, crude oil has long been transported in bulk on the waterways, and thus the nationâs inland and coastal barge operators have considerable familiarity with the commodityâs hazards, transportation risks, and potential emergency scenarios. Their safety assurance system was tested less by the introduction of an unfamiliar commodity and more by the large increase in oil volumes moving on waterways that are inherently sensitive to spills. 3 National Transportation Safety Board, âCollision between Bulk Carrier Summer Wind and the Miss Susan Tow, Houston Ship Channel, Lower Galveston Bay, Texas,â Accident Report NTSB/MAR-15/01 (Washington, D.C., March 22, 2014), https://www.ntsb.gov/investigations/ AccidentReports/Pages/MAR1501.aspx.
70 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES The same commodity familiarity existed in the pipeline industry, in which carriers operate systems that are purpose-built to move hazardous liquids and gases over long distances, under high pressure and in high vol- umes. While the growth in production of domestic crude oil and natural gas has led to the construction of more pipeline capacity and to changes in the direction, commodity mix, and volumes of some transmission lines, the safety challenges arising from this energy liquids and gas traffic were by no means new to the industry. What has been new, however, is the expansion of oil and gas production to regions of the country that had limited pipeline capacity because they did not have a large production role historically. This chapter reviews the safety records of each of the three long- distance modes when transporting crude oil, ethanol, and natural gas. Consideration is given to the number, severity, and causes of incidents re- ported to regulators over the past decade and to safety issues that have been identified during investigations of severe incidents. In the case of pipelines, incidents are reported to and investigated by the Pipeline and Hazardous Materials Safety Administration (PHMSA). PHMSA is the main federal agency responsible for overseeing the safe transportation of hazardous materials in all modes, not only pipelines. However, in the case of rail and water transportation, PHMSA shares hazardous materials safety oversight responsibility with the Federal Railroad Administration (FRA) and the U.S. Coast Guard (USCG), respectively. Incident data and incident inves- tigation results from the three agencies are reviewed, along with relevant findings and recommendations from the National Transportation Safety Board (NTSB), which conducts detailed investigations of major hazardous materials incidents in all modes. The chapter concludes with a review of major safety-related policies that have been put in place since the domestic energy revolution, par- ticularly for rail transportation, which had not previously carried large quantities of oil or ethanol. The rapid and unanticipated increase in flam- mable liquids traffic moving over the nationâs rail network caught some traversed communities, especially in rural areas, off guard and increased concern about their capacity to respond to a large-scale incident. Policy re- sponses to these concerns, including changes in regulations and initiatives to strengthen state and local emergency response preparations, are discussed. PIPELINE SAFETY TRENDS Crude Oil Pipelines PHMSA requires hazardous liquid pipeline operators to report unintended releases that meet certain significance thresholds based on release quantity and impact severity. Specifically, operators must report any unintended
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 71 release that involves 5 gallons or more and/or an explosion, fire, serious in- jury, or significant (i.e., $50,000 or more) property loss and cleanup costs.4 Releases from any component of the pipeline facility, including line pipe, tanks, valves, manifolds, and pumps, must be reported. The operator must report the proximate cause and the consequences of the incident.5 Figure 4-1 shows the number of incidents of reported crude oil releases from 2005 to 2015. From 2005 until 2011, the number was consistently between 143 and 166 per year. After 2011, incidents increased steadily, up by nearly 50 percent when comparing 2015 to 2005. However, aggregate release volumes have fluctuated widely from year to year, with a down- ward secular trend. Because of Hurricane Katrina, release volumes in 2005 reached a record 100,000 barrels, including nearly 50,000 barrels spilled in the aftermath of the storm. The 2011â2015 period coincided with the build out of the pipeline network to accommodate the increasing volumes of crude oil produced by hydraulic fracturing as well as Canadian imports. As discussed in the previous chapter, the crude oil pipeline network grew by 54 percent from 2005 to 2015. Accordingly, when controlling for the increase in total system miles, the incident rate normalized by pipeline miles (i.e., releases per pipeline mile) remained steady over the decade. Because major releases are a primary concern, and small releases can distort the safety picture, PHMSA distinguishes âsignificantâ releases as those that involve an explosion, fire, serious injury, or significant property damage and/or when 50 gallons or more of crude oil is released or 5 gallons or more is released into water. If one focuses on these incidentsâthat is, those that have accounted for more than 98 percent of released quantitiesâ the upward trend in reported incidents persists (see Figure 4-2). However, when normalizing for traffic volume (i.e., barrel-miles), the increase in sig- nificant incidents after 2010 appears to be less pronounced and the volume of transported crude oil released has declined, as shown in Figure 4-3. PHMSA data on the proximate cause of incidents suggest that a reason for the recent annual increase in incidents is more operator reports of equip- ment failures. Until 2009, corrosion had been the main cause of releases. However, during 2010 to 2015, corrosion-related releases remained fairly steady, as did most other causes (see Figure 4-4). A notable exception was equipment-related releases, which overtook corrosion as the main cause of incidents after 2012. According to PHMSA, equipment failures usually result in a release that is contained on company property, rarely causing 4 49 CFR 195.50. 5 The PHMSA incident database is publicly accessible online at https://www.phmsa.dot.gov/ pipeline/library/data-stats/pipelineincidenttrends.
72 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES FIGURE 4-1 Crude oil pipeline releases reported to PHMSA in the United States, 2005â2015. SOURCE: PHMSA, âPipeline Incident Flagged Files,â accessed November 23, 2016, http://www.phmsa.dot.gov/pipeline/library/data-stats/flagged-data-files. FIGURE 4-2 Significant crude oil pipeline releases reported to PHMSA in the United States, 2005â2015. SOURCE: PHMSA, âPipeline Incident Flagged Files,â accessed November 23, 2016, http://www.phmsa.dot.gov/pipeline/library/data-stats/flagged-data-files. 4-1 0 20,000 40,000 60,000 80,000 100,000 120,000 0 50 100 150 200 250 300 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 All incidents Incidents Release Volume Barrels released 4-2 0 20,000 40,000 60,000 80,000 100,000 120,000 0 10 20 30 40 50 60 70 80 90 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Significant incidents Significant incidents Release volume Barrels released
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 73 FIGURE 4-3 Significant incidents and release volumes per trillion barrel-miles of crude oil transported in the United States, 2005â2015. SOURCE: PHMSA, âPipeline Incident Flagged Files,â accessed November 23, 2016, http://www.phmsa.dot.gov/pipeline/library/data-stats/flagged-data-files. 4-3 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 0 5 10 15 20 25 30 35 40 45 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Barrels released per trillion barrel-miles Significant incidents Barrels released Significant incidents per trillion barrel-miles FIGURE 4-4 Crude oil pipeline incidents by reported cause in the United States, 2010â2015. NOTE: For causes that rely on specialized terminology, the following definitions ap- ply: âequipmentâ indicates a failure of control systems or relief equipment; âincor- rect operationâ means operator error; ânatural forceâ refers to damage stemming from naturally occurring phenomena, for instance, earthquakes and floods; and âoutside forceâ indicates damage from human actions other than excavation, such as accidental vehicle collision and earth-moving activity. SOURCE: PHMSA. 4-4 0 20 40 60 80 100 120 2010 2011 2012 2013 2014 2015 CORROSION EQUIPMENT EXCAVATION DAMAGE INCORRECT OPERATION MATERIAL PIPE/WELD FAILURES NATURAL FORCE OTHER OUTSIDE FORCE All incidents
74 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES injury or significant damage.6 Common equipment-related incidents are small releases from faulty seals and gaskets or storage tanks overfilling as a result of meters malfunctioning. The increase in reports of equipment failures explains the seemingly contrary trends of more total incidents but lower total release volumes. Another important reason for inconsistencies in annual release numbers and total release quantities is that one or two very large incidents can have a disproportionate influence on release totals. This influence is evident in 2010, when more than half of the 52,000 barrels of crude oil was released from two incidents: a 20,000-barrel release in Marshall, Michigan, and 7,500-barrel release in Romeoville, Illinois. Likewise, in 2008, 31,000 of the more than 59,000 barrels of crude oil spilled that year were from a single pipeline release in Denver City, Texas. More discussion is given to such severe incidents later in the section on pipelines. As might be expected, states with the most pipeline mileage are most likely to experience a release; indeed, California, Oklahoma, and Texas are the most frequent locations for significant incidents, as shown in Figure 4-5. It is notable that several states that have experienced some of the largest gains in crude oil pipeline mileage since the start of the domestic energy revolution are now among states experiencing the most incidents. Three of the top five states in added crude oil pipeline mileage are New Mexico, North Dakota, and Wyoming (see Table 4-1). They are now among the top 10 states in significant releases reported since 2010 (see Figure 4-5). As industry met the growing demand for additional transmission pipe- line infrastructure, for crude oil and the other petroleum liquid and gas commodities, PHMSA inspections began registering critical problems with pipeline quality. During the 2008 and 2009 construction seasons, its inspec- tors observed new pipelines failing in field pressure tests and from welds,7 as well as substandard metallurgical properties in pipe materials that ex- hibited significant weakness.8 The regulator issued advisory bulletins on these findings and held workshops in which the public, industry, and other 6 Pipeline and Hazardous Materials Safety Administration, âFact Sheet: Stakeholder Com- municationsâEquipment Failure,â accessed September 19, 2017, https://primis.phmsa.dot. gov/comm/FactSheets/FSEquipmentFailure.htm. 7 Pipeline and Hazardous Materials Safety Administration, âPipeline Safety: Girth Weld Quality Issues Due to Improper Transitioning, Misalignment, and Welding Practices of Large Diameter Line Pipe (ABD-10-03),â PHMSA-Advisory Bulletins, March 24, 2010, https://www. gpo.gov/fdsys/pkg/FR-2010-03-24/html/2010-6528.htm. 8 Pipeline and Hazardous Materials Safety Administration, âPipeline Safety: Potential Low and Variable Yield and Tensile Strength and Chemical Composition Properties in High Strength Line Pipe (ABD-09-01),â PHMSA-Advisory Bulletins, May 21, 2009, https://www. gpo.gov/fdsys/pkg/FR-2009-05-21/html/E9-11815.htm.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 75 TABLE 4-1 Top Five States with Most New Miles of Crude Oil Pipelines, 2010â2015 State Pipeline Miles Added, 2010â2015 Percentage Increase in Mileage in Each State Percentage Share of All New Pipeline Miles Added in United States Texas 8,138 64% 44% Oklahoma 2,801 78% 15% North Dakota 1,535 101% 8% Wyoming 842 25% 5% New Mexico 591 52% 3% Subtotal 13,906 NA 75% U.S. Total 18,451 NA 100% SOURCE: PHMSA, âPipeline Miles and Facilities 2010+.â FIGURE 4-5 Reports of significant releases of crude oil from pipelines by state, 2010â2016. * Incident reports filed as of September 14, 2016. Mileage data were not avail- able for 2016. SOURCES: PHMSA, âPipeline Incident Flagged Files.â PHMSA, âPipeline Miles and Facilities 2010+.â - 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 0 10 20 30 40 50 60 70 80 90 2010 2011 2012 2013 2014 2015 2016* TX OK CA ND KS LA IL WY MN NM All other states Pipeline Miles Significant releases Total pipeline miles
76 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES regulatory partners could learn about the issues and find solutions.9 Efforts jointly made by the regulator and the industry to assure safety early in the pipeline lifecycle are intended to mitigate the risk of threats to integrity management that can worsen over time. It is important to emphasize, however, that severe pipeline incidents are low-frequency events that occur most often because of time-dependent mechanisms such as corrosion and stress cracking. It is questionable, there- fore, that the pipeline system would experience an increase in the occur- rence of time-dependent failures because of new pipeline capacity added during the past decade. New crude oil pipelines will undoubtedly suffer eventually from some of the same failure mechanisms observed in the mature system. More pipeline milesâand thus more exposureâcan be ex- pected to increase the number of pipeline failures, all else being equal. At the same time, however, the mature, national system contains tens of thou- sands of miles of pre-1970 pipelines.10 The newer pipelines built to accom- modate oil produced by hydraulic fracturing will have the benefit of more advanced materials and construction techniques, state-of-the-art corrosion protections such as fusion-bonded epoxy coatings, and components and designs that enable faster leak detection and easier cleaning, non intrusive monitoring, and internal inspections. Given the superior technology, the safety performance of the new pipelines may be substantially better than that of the existing system. Of course, these pipelines connect with legacy trunklines, some of which had previously carried crude oil from the Gulf Coast to the Midwest but have since had their flow directions reversed. In some cases, the pipelines may have been repurposed to carry crude oil after carrying other products, including gas. Accordingly, the challenge that persists is ensuring that the mature pipeline network, much of it decades old, continues to perform safely, irrespective of whether it is used to carry product originating in the new oil- and gas-producing regions or elsewhere. Inasmuch as changes in the amount of product carried or in the flow direction of a pipeline can require modifications to components such as valves and pumps and create new pressure profiles and stress demands, there remains the possibility of risks being introduced into the legacy network because of the increase in oil production from hydraulic fracturing. PHMSA has issued guidance to op- 9 Pipeline and Hazardous Materials Safety Administration, âPipeline Construction,â Pipe- line Technical Resources, https://primis.phmsa.dot.gov/construction. 10 Pipeline and Hazardous Materials Safety Administration, âThe State of the National Pipeline Infrastructure,â 7, accessed September 19, 2017, https://opsweb.phmsa.dot.gov/ pipelineforum docs/Secretarys%20Infrastructure%20Report_Revised%20per%20PHC_103111.pdf. Accord- ing to PHMSA data from 2010, more than half of the mileage on the hazardous liquids pipeline network was built before 1970; https://opsweb.phmsa.dot.gov/pipelineforum/docs/ Secretarys%20Infrastructure%20Report_Revised%20per%20PHC_103111.pdf.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 77 erators for avoiding and controlling such risks, which may not be detected in incident data for many years.11 While PHMSA has begun to track these reversals and conversions, its pipeline incident database omits information on whether incidents involved pipelines that had been converted from car- rying other products or had their flow reversed. NGL Pipelines Inasmuch as the hydraulic fracturing revolution has led to more natural gas liquid (NGL) shipments, the challenge for industry and regulators is to ensure that the new and existing pipelines operate safely. There is no practical and economic alternative to moving highly volatile NGLs in bulk quantities by means other than pipeline. Notably, unlike crude oil, ethane must be moved by pipeline because of its high-pressure requirements. Ac- cordingly, the barrel volumes in an unintentional release can be quite large because the escaping liquids are quickly volatilized. NGL products can pose a risk of flammability and asphyxiation, and a large pressurized release may ignite and result in a highly energetic explosion and fireball. Figure 4-6 shows the trend in significant NGL pipeline releases reported to PHMSA from 2005 to 2015. As with crude oil, there has been a slight upward trend in incidents since 2005, but with substantial year-to-year variability both in incidents and in total release volumes. This same pattern of year-to-year variability is evident when controlling for traffic volumes (see Figure 4-7). Based on these early incident data at least, there is no indication that the growing volumes of NGLs moved by pipeline have cre- ated a new safety risk. As noted above, however, the factors that can cause incidents, including corrosion, can be time-dependent, complicating efforts to assess safety performance over a relatively short period. Natural Gas Pipelines As discussed in Chapter 3, natural gas is a much more homogeneous and fungible commodity than crude oil and NGLs. In this sense, once natural gas enters the transmission system it integrates with other natural gas produced domestically, making it impractical to distinguish between gas shipments originating in the new producing regions versus other production areas. The annual number of significant natural gas releases fluctuated be- tween 40 and 60 per year from 2005 to 2015 (see Figure 4-8). After 2010, PHMSA began to track release volumes. From 2010 to 2015, there was a 11 Pipeline and Hazardous Materials Safety Administration, âGuidance for Pipeline Flow Reversals, Product Changes, and Conversion to Service,â September 2014, http://www. occeweb.com/PLS/2014Gas/Guide-Flo%20Rev-Prod%20Ch-Conver.pdf.
78 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES FIGURE 4-7 Significant incidents and release volumes per trillion barrel miles of NGLs transported in the United States, 2005â2015. NOTE: Data are from PHMSAâs highly volatile liquids (HVL) incident files but exclude ammonia, refined products, and carbon dioxide pipelines. SOURCE: PHMSA, âPipeline Incident Flagged Files.â FIGURE 4-6 Significant NGL pipeline releases reported to PHMSA in the United States, 2005â2015. NOTE: Data are from PHMSAâs highly volatile liquids (HVL) incident files but exclude ammonia, refined products, and carbon dioxide pipelines. SOURCE: PHMSA, âPipeline Incident Flagged Files.â 4-7 0 20,000 40,000 60,000 80,000 100,000 120,000 0 10 20 30 40 50 60 70 80 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Barrels released per trillion barrel-miles Significant incidents Total volume released Incidents per trillion barrel-miles 4-6 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 0 5 10 15 20 25 30 35 40 45 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Incidents Significant incidents Total volume released Volume released, bbl
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 79 consistent decline in the total amount of gas released in incidents. However, when normalized for shipment volumes, there appears to have been a steady increase in the incident rate (see Figure 4-9), which can be attributed in part to a decline in the amount of natural gas shipments during that period, as discussed earlier in Chapter 3. As discussed in the next section, several large releases contributed to this pattern. For example, two incidents in Colorado and one in Texas during 2010 led to the release of more than 770,000 cubic feet of gas, and two incidents in California and Florida during 2012 led to the release of about 880,000 cubic feet.12 Notable Major Pipeline Incidents Table 4-2 lists incidents of major pipeline releases involving crude oil, NGLs, and natural gas from 2010 to 2015. In addition to large release volumes, the consequences of some of these incidents include loss of life, injuries, evacuations, environmental harm, and property damage. Reported 12 Pipeline and Hazardous Materials Safety Administration, âPipeline Incident Flagged Files.â Data are from PHMSAâs HVL incident files but exclude ammonia, refined products, and carbon dioxide pipelines. 4-8 0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 0 10 20 30 40 50 60 70 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Significant incidents Significant incidents Total volume released Volume released, billions of cubic feet FIGURE 4-8 Significant natural gas pipeline releases reported to PHMSA in the United States, 2005â2015. NOTE: Natural gas volumes released were not recorded before 2010. SOURCE: PHMSA, âPipeline Incident Flagged Files.â
80 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES FIGURE 4-9 Significant natural gas pipeline incidents and volumes released per billion cubic feet reported to PHMSA in the United States, 2010â2015. NOTES: Natural gas release volumes were not recorded before 2010. Normalized according to natural gas traffic volumes reported by EIA. Bcf = billions of cubic feet; mmcf = millions of cubic feet. SOURCE: EIA, âU.S. International and Interstate Movements of Natural Gas by State,â accessed November 8, 2016, http://www.eia.gov/dnav/ng/ng_move_ist_ a2dcu_nus_a.htm. proximate causes vary, but include corrosion damage, excavation damage aided by a lack of pipeline markers, and poor performance of supervisory control systems to detect and contain leaks. Many of the incidents were investigated by NTSB. In the case of the 2010 San Bruno natural gas transmission pipeline explosion, in which 8 people died and 108 homes were destroyed or damaged, NTSB concluded that the pipeline operator had not properly followed its integrity management program, as required by regulation.13 The agency raised similar concerns after investigating the 2012 natural gas pipeline incident in Sissonville, West Virginia. The 2010 Marshall, Michigan, crude oil release was found to have been caused by corrosion fatigue cracking and made worse by the pipeline operator con- tinuing to pump product through the ruptured line after misinterpreting pressure loss data.14 13 National Transportation Safety Board, âPacific Gas and Electric Company Natural Gas Transmission Pipeline Rupture and Fire,â August 30, 2011, http://ntsb.gov/investigations/ AccidentReports/Pages/PAR1101.aspx. 14 National Transportation Safety Board, âEnbridge Incorporated Hazardous Liquid Pipe- line Rupture and Release,â July 10, 2012, http://ntsb.gov/investigations/AccidentReports/ Pages/PAR1201.aspx. 4-9 0.000 0.005 0.010 0.015 0.020 0.025 0.0000 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009 0.0010 0.0011 2010 2011 2012 2013 2014 2015 Incidents per bcf Significant incidents per billion cubic feet Total volume released Mmcf released per bcf transported
81 T A B L E 4 -2 U .S . C ru de O il, N at ur al G as , an d N G L P ip el in e In ci de nt s w it h Se ve re C on se qu en ce s, 2 01 0â 20 15 L oc at io n M on th /Y ea r Pr od uc t V ol um e (l iq ui ds = b bl ; ga s = m cf ) Fa ta lit y In ju ry Fi re E va cu at io n E nv ir on m en ta l/ Pr op er ty D am ag e C le bu rn e, T X 6/ 20 10 N at ur al G as 17 2, 00 0 1 6 Y N Y E ag le C ou nt y, C O 6/ 20 10 N at ur al G as 25 0, 00 0 N N N Y N Se al y, T X 7/ 20 10 N at ur al G as 20 8, 45 8 N 1 Y Y Y Sa n B ru no , C A 9/ 20 10 N at ur al G as 47 ,6 00 8 51 Y Y Y A rv in , C A 5/ 20 12 N at ur al G as 58 5, 45 7 N N N N Y Si ss on vi lle , W V 12 /2 01 2 N at ur al G as 76 ,0 00 N N Y Y Y K it C ar so n, C O 12 /2 01 2 N at ur al G as 31 3, 87 0 N N N N N M el bo ur ne , FL 12 /2 01 2 N at ur al G as 29 3, 97 6 N N N N Y Po nd C re ek , O K 2/ 20 10 N G L 13 ,7 18 N N N Y Y L at an , O K 2/ 20 10 N G L 12 ,8 36 N N Y N N L af ou rc he , L A 2/ 20 13 N G L 23 ,7 02 1 N Y N N E ri e, I L 8/ 20 13 N G L 18 ,4 00 N N Y Y Y L it tl et on , W V 8/ 20 13 N G L 11 ,4 05 N N N N N C ol lie rs , W V 1/ 20 15 N G L 30 ,5 65 N N Y Y Y Ja ck so n C ou nt y, T X 1/ 20 15 N G L 27 ,1 23 N N N N Y M ar sh al l, M I 7/ 20 10 C ru de O il 20 ,0 82 N N N Y Y R om eo vi lle , IL 9/ 20 10 C ru de O il 6, 43 0 N N N Y Y L ev el la nd , T X 10 /2 01 0 C ru de O il 10 ,2 00 N N N N N Io la , T X 1/ 20 11 C ru de O il 6, 91 1 N N N N Y C hi co , T X 6/ 20 11 C ru de O il 12 ,2 29 N N N N N M ag no lia , A R 3/ 20 13 C ru de O il 5, 60 0 N N N N Y M ou nt ra il C ou nt y, N D 9/ 20 13 C ru de O il 20 ,6 00 N N N N Y M oo ri ng sp or t, L A 10 /2 01 4 C ru de O il 4, 50 9 N N N Y Y SO U R C E : PH M SA , âP ip el in e In ci de nt F la gg ed F ile s. â
82 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES PHMSA has sought to improve the ways that pipeline operators imple- ment requirements for integrity management programs to prevent incidents like these. As early as 2002 when it first expanded the scope of its integrity management program regulations,15 and again in 2016,16 the agency has articulated the importance of continuous improvement in the risk model- ing and the risk-assessment capabilities of operators. In observing a lack of progress by industry in the development of these capabilities, PHMSA proposed regulations in 2016 that would prescribe several functional goals that risk assessments should be able to achieve, including validation of risk models with historical information and the identification of interactive threats (i.e., when multiple concurrent threats act on one another) and un- certainty in the models and data.17 To further encourage industry to make improvements along these lines, PHMSA has convened events to share best practices for risk modeling and assessment and sponsored research to pro- vide technical assistance to the industry. The agency held a forum in 2011 on improving pipeline risk assessments and a workshop on risk-modeling methodologies in 2015.18 It published three reports in 2016 and 2017 on pipeline risk model guidelines based on the probabilistic quantitative risk analysis approach, a risk management literature review and industry survey of risk management decision making, and a review of approaches to imple- menting risk-modeling and integrity-management programs.19 RAIL SAFETY TRENDS Federal regulations governing hazardous materials transportation require that certain types of hazardous materials incidents be reported by the car- rier to PHMSA.20 A report must be filed if there is an unintentional release during any transportation phase, including loading and unloading, and/or a 15 Pipeline and Hazardous Materials Safety Administration, âPipeline Safety: Pipeline Integ- rity Management in High-Consequence Areas (Hazardous Liquid Operators with Less Than 500 Miles of Pipelines),â 67 FR 2136 (2002), 2143, https://primis.phmsa.dot.gov/iim/docsr/ smallIMPrulefinal.pdf. 16 Pipeline and Hazardous Materials Safety Administration, âPipeline Safety: Safety of Gas Transmission and Gathering Pipelines,â 81 FR 20722 (2016), 20763â64, https://www. regulations.gov/contentStreamer?documentId=PHMSA-2011-0023-0118&disposition=attach ment&contentType=pdf. 17 Ibid. 18 Pipeline and Hazardous Materials Safety Administration, âPHMSA Meeting Registra- tion and Document Commenting,â PHMSA Public Meetings and Documents, https://primis. phmsa.dot.gov/meetings. 19 Pipeline and Hazardous Materials Safety Administration, âFinal Reports,â Research and Development Program, https://primis.phmsa.dot.gov/matrix/FinalReports.rdm. 20 49 CFR 171 Subpart BâIncident Reporting, Notification, BOE Approvals and Authorization.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 83 tank car that contains any amount of hazardous materials suffers structural damage regardless of whether there was a release. Accordingly, PHMSAâs incident data files should contain records of all instances in which tank cars carrying crude oil and ethanol have experienced unintentional releases and/ or been damaged during a crash, such as a derailment or yard collision. The reports contain, among other things, the quantity released, the compo- nent of the tank car that failed (e.g., tank shell, valve, bottom outlet), cause of the failure (e.g., derailment, unsecured closure, faulty fitting), tank car design type, speed of the train or car being moved, and the consequences (e.g., fire, evacuation, injuries, property damage). Incidents with serious consequences will usually be investigated by the Federal Railroad Associa- tion (FRA) and in some cases by NTSB in more detail. Figures 4-10 and 4-11 show the total number of incidents involving tank car shipments of crude oil and ethanol reported from 2005 to 2015. Ethanol shipments were involved in 968 total incidents and crude oil ship- ments in 448. This sizable difference in incidents by commodity type is at least partially attributable to the fact that ethanol has been transported in large volumes by rail since about 2006, compared with crude oilâs emer- gence as a major rail traffic segment after 2010. In both cases, however, the large majority of reported incidents is small releases caused by improperly secured closures or component failures such as a deteriorated gasket or pressure seal. They are usually detected during inspection while a tank is stationary on a siding or at a rail yard. Most involve vapor quantities or fractions of a gallon. For example, of the 427 reports of crude oil releases from stationary tanks cars, 209 had leaks of less than 1 gallon. Even small leaks from a stationary tank car at a siding or a yard can in- cur costs, such as service disruptions while the car is being isolated and the leak contained. The focus here, however, is on the approximately 5 percent of incidents that occur when a tank car is damaged while moving, usually because of a derailment or a yard collision (e.g., during coupling). These incidents tend to involve larger releases and more severe consequences. From 2005 to 2015, there were 58 such incidents involving ethanol ship- ments and 21 involving crude oil shipments (see Figures 4-12 and 4-13 and Tables 4-3 and 4-4). These 79 incidents accounted for most of the product lost and all severe consequences such as fires, evacuations, and injuries. The 21 crude oil incidents accounted for more than 98 percent of the total 1.7 million gallons (40,400 barrels) of crude oil released unintentionally from tank cars during the period. Likewise, the 58 ethanol incidents ac- counted for more than 93 percent of the 2.8 million gallons (66,900 barrels) of ethanol released. The relatively small number of incidents makes it difficult to glean patterns, such as their geographic distribution. It is notable, however, that more than half of the contiguous United States have experienced ethanol
84 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES FIGURE 4-10 Reports of unintentional hazardous materials releases or damaged tank cars involving ethanol shipments by rail in the United States, 2005â2015. SOURCE: PHMSA, âIncident Statistics,â accessed November 8, 2016, http://www. phmsa.dot.gov/hazmat/library/data-stats/incidents. FIGURE 4-11 Reports of unintentional hazardous materials releases or damaged tank cars involving crude oil shipments by rail in the United States, 2005â2015. SOURCE: PHMSA, âIncident Statistics.â 4-10 910 reports in which tank cars were stationary and/or no crash or derailment (94%) 58 reports in which tank cars were moving and/or crash or derailment (6%) Ethanol rail incidents, 2005â2015 4-11 427 reports in which tank cars were stationary and/or no crash or derailment (95%) 21 reports in which tank cars were moving and/or crash or derailment (5%) Crude oil rail incidents, 2005â2015
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 85 FIGURE 4-12 Ethanol tank car derailments and collisions and release quantities in the United States, 2005â2015. SOURCE: PHMSA, âIncident Statistics.â FIGURE 4-13 Crude oil tank car derailments and collisions and release quantities in the United States, 2005â2015. SOURCE: PHMSA, âIncident Statistics.â 0 1 2 3 4 5 6 7 8 9 10 11 12 0 2 4 6 8 10 12 14 16 18 20 22 24 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Barrels Released Incidents Barrels (thousands) Incidents 0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Barrels (thousands) Barrels Released Incidents Incidents
86 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES TABLE 4-3 Tank Car Derailments and Collisions Involving the Unintentional Release of Ethanol in the United States, 2005â2015 Year Incidents Cars Incurring Damage or Release Release (gallons) Release (barrels) 2015 4 17 102,934 2,451 2014 2 2 4,025 96 2013 2 4 4,581 109 2012 8 25 340,589 8,109 2011 8 40 1,084,338 25,818 2010 5 13 81,806 1,948 2009 11 28 332,560 7,918 2008 1 1 12,447 296 2007 6 9 108,506 2,583 2006 7 27 525,223 12,505 2005 4 5 22,580 538 Total 58 171 2,619,588 62,371 SOURCE: PHMSA, âIncident Statistics.â TABLE 4-4 Tank Car Derailments and Collisions Involving the Unintentional Release of Crude Oil in the United States, 2005â2015 Year Incidents Cars Incurring Damage or Release Release (gallons) Release (barrels) 2015 7 58 598,183 14,242 2014 3 22 47,149 1,123 2013 5 47 940,457 22,392 2012 0 0 0 0 2011 1 1 1 <1 2010 1 1 4,800 114 2009 1 1 1 <1 2008 2 6 80,886 1,925 2007 0 0 0 0 2006 0 0 0 0 2005 1 1 1,500 36 Total 21 137 1,672,976 39,833 SOURCE: PHMSA, âIncident Statistics.â
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 87 incidents. Because ethanol is blended in gasoline in all states, it is shipped over a larger percentage of the countryâs railroad network. Because ship- ments originate in corn-producing regions, the Midwest is the most com- mon site of the 58 incidents, with more than one-third occurring in Illinois, Indiana, Iowa, and Ohio. As might be expected given their proximity to the Bakken formation, North Dakota and Montana have experienced the most crude oil train derailments, but otherwise there is no apparent pattern to the 21 incidents. Notable Major Railroad Incidents Table 4-5 lists the 20 ethanol and crude oil incidents that have had the most severe consequences of the 79, many involving fires, public evacuations, and thousands of barrels of lost product. As the timeline of incidents indicates, until 2013 nearly all large flammable liquids incidents involved ethanol shipments, including the earliest dating to October 2006, when 23 cars of an 83-car unit train derailed in New Brighton, Pennsylvania. Following an investigation of that incident, NTSB concluded the cause was a rail fracture from a defect that had gone undetected by railroad inspectors (see Table 4-5). Of the next six major ethanol derailments, from Painesville, Ohio, in October 2007 to Columbus, Ohio, in July 2012, all were determined to have been caused by track, roadbed, or structure failures, predominantly worn and broken rails.21 Indeed, all of these notable incidents have in- volved derailments. Moreover, track infrastructure failures have overwhelmingly been the cause for these derailments. A review of FRA safety data indicates that track, roadbed, and structure problems caused more than 44 percent of all report- able derailments from 2005 to 2015.22 In the case of these severe crude oil and ethanol derailments, track flaws caused about 80 percent of the derail- ments. As shown in Table 4-5, 16 incidents involved track, roadbed, and structure failures, while the rest of the incidents were the result of equipment failure, human error, or causes unknown at this time. It merits noting that the incidents at Mount Carbon, West Virginia; Lynchburg, Virginia; and Mosier, Oregon, involved derailments that occurred shortly after the track was inspected, which raised questions about updated track safety regulations issued by FRA in 2014, which are discussed later in this chapter. 21 National Transportation Safety Board, âDerailment of CN Freight Train U70691-18 with Subsequent Hazardous Materials Release and Fire, Cherry Valley, Illinois June 19, 2009,â Ac- cident Report NTSB/RAR-12/01 (Washington, D.C., February 14, 2012), https://www.ntsb. gov/investigations/AccidentReports/Pages/RAR1201.aspx. 22 U.S. Federal Railroad Administration, âFederal Railroad Administration Office of Safety Analysis: 3.10-Accident Causes,â accessed November 23, 2016, http://safetydata.fra.dot.gov/ officeofsafety/publicsite/Query/inccaus.aspx.
88 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES TABLE 4-5 Notable Major Ethanol and Crude Oil Tank Car Incidents in the United States, 2005â2015 Location Month and Year Tank Cars Derailed Tank Cars with Release MPH Product Gallons Barrels Fatality Injury Fire Evacuation Environmental Damage New Brighton, PAâ¡ 10/06 23 20 37 Ethanol 485,278 11,554 0 0 Yes Yes Yes Painesville, OHâ 10/07 7 4 48 Ethanol 55,200 1,314 0 0 Yes Yes No Cherry Valley, IL* 6/09 19 15 37 Ethanol 323,963 7,713 1 8 Yes Yes No Arcadia, OHâ 2/11 31 31 46 Ethanol 834,840 19,877 0 0 Yes Yes Yes Tiskilwa, ILâ 10/11 10 9 34 Ethanol 162,000 3,857 0 0 Yes Yes No Columbus, OHâ¡ 7/12 3 3 23 Ethanol 54,748 1,304 0 2 Yes Yes Yes Plevna, MTâ 8/12 17 12 25 Ethanol 245,336 5,841 0 0 Yes No No Lesterville, SDâ¡ 9/15 7 3 10 Ethanol 49,748 1,184 0 0 Yes No No Alma,WIh 11/15 15 5 26 Ethanol 20,413 486 0 0 No Yes Yes Luther, OK* 8/08 8 5 34 Crude Oil 80,746 1,923 0 0 Yes Yes No Parkers Prairie, MNâ¡ 03/13 14 3 40 Crude Oil 30,000 714 0 0 No No Yes Aliceville, ALâ 11/13 26 25 39 Crude Oil 455,520 10,846 0 0 Yes No Yes Casselton, NDÂ¤ 12/13 20 18 42 Crude Oil 474,936 11,308 0 0 Yes Yes No New Augusta, MSâ 01/14 15 7 47 Crude Oil 90,000 2,143 0 0 No Yes No Vandergrift, PA* 2/14 19 4 31 Crude Oil 10,000 238 0 0 No Yes Yes Lynchburg, VAâ 4/14 17 1 24 Crude Oil 29,416 700 0 0 Yes Yes Yes Mount Carbon, WVâ¡ 2/15 27 20 33 Crude Oil 378,034 9,001 0 1 Yes Yes Yes Galena, ILw 3/15 21 10 23 Crude Oil 110,543 2,632 0 0 Yes Yes No Heimdall, NDu 5/15 6 5 24 Crude Oil 98,090 2,335 0 0 Yes Yes No Culbertson, NDâ 7/15 22 5 44 Crude Oil 27,201 648 0 0 No Yes No NOTE: â¡ = internal rail defect; â = other rail defect; * = track or roadbed problem such as inadequate storm water management or wide gauge; Â¤ = broken axle; h = human error; w = wheel defect; u = under investigation. SOURCE: PHMSA, âIncident Statistics.â
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 89 TABLE 4-5 Notable Major Ethanol and Crude Oil Tank Car Incidents in the United States, 2005â2015 Location Month and Year Tank Cars Derailed Tank Cars with Release MPH Product Gallons Barrels Fatality Injury Fire Evacuation Environmental Damage New Brighton, PAâ¡ 10/06 23 20 37 Ethanol 485,278 11,554 0 0 Yes Yes Yes Painesville, OHâ 10/07 7 4 48 Ethanol 55,200 1,314 0 0 Yes Yes No Cherry Valley, IL* 6/09 19 15 37 Ethanol 323,963 7,713 1 8 Yes Yes No Arcadia, OHâ 2/11 31 31 46 Ethanol 834,840 19,877 0 0 Yes Yes Yes Tiskilwa, ILâ 10/11 10 9 34 Ethanol 162,000 3,857 0 0 Yes Yes No Columbus, OHâ¡ 7/12 3 3 23 Ethanol 54,748 1,304 0 2 Yes Yes Yes Plevna, MTâ 8/12 17 12 25 Ethanol 245,336 5,841 0 0 Yes No No Lesterville, SDâ¡ 9/15 7 3 10 Ethanol 49,748 1,184 0 0 Yes No No Alma,WIh 11/15 15 5 26 Ethanol 20,413 486 0 0 No Yes Yes Luther, OK* 8/08 8 5 34 Crude Oil 80,746 1,923 0 0 Yes Yes No Parkers Prairie, MNâ¡ 03/13 14 3 40 Crude Oil 30,000 714 0 0 No No Yes Aliceville, ALâ 11/13 26 25 39 Crude Oil 455,520 10,846 0 0 Yes No Yes Casselton, NDÂ¤ 12/13 20 18 42 Crude Oil 474,936 11,308 0 0 Yes Yes No New Augusta, MSâ 01/14 15 7 47 Crude Oil 90,000 2,143 0 0 No Yes No Vandergrift, PA* 2/14 19 4 31 Crude Oil 10,000 238 0 0 No Yes Yes Lynchburg, VAâ 4/14 17 1 24 Crude Oil 29,416 700 0 0 Yes Yes Yes Mount Carbon, WVâ¡ 2/15 27 20 33 Crude Oil 378,034 9,001 0 1 Yes Yes Yes Galena, ILw 3/15 21 10 23 Crude Oil 110,543 2,632 0 0 Yes Yes No Heimdall, NDu 5/15 6 5 24 Crude Oil 98,090 2,335 0 0 Yes Yes No Culbertson, NDâ 7/15 22 5 44 Crude Oil 27,201 648 0 0 No Yes No NOTE: â¡ = internal rail defect; â = other rail defect; * = track or roadbed problem such as inadequate storm water management or wide gauge; Â¤ = broken axle; h = human error; w = wheel defect; u = under investigation. SOURCE: PHMSA, âIncident Statistics.â
90 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES Examinations of the post-crash factors that contributed to the severity of these derailments have often pointed to problems with the integrity of the tank cars used. In reporting on each of the ethanol incidents, NTSB observed that the derailed DOT-111 tank cars had exhibited poor crash- worthiness and thermal resistance. In the 2006 New Brighton incident, heat from the ensuing fire caused a tank car to overpressurize and rupture. While the agency did not recommend changes to the ethanol tank car fleet following this early incident, follow-on investigations continued to point to shortcomings in the crash performance of this general service car. Of the 10 ethanol cars that derailed in Tiskilwa, Illinois, in 2011, 9 were severely damaged, with most sustaining head and shell punctures and 3 experienc- ing breached tanks as a result of exposure to the high temperatures of a pool fire. Following this incident, NTSB concluded that âDOT-111 tank cars are inadequately designed to prevent punctures and breaches and that catastrophic release of hazardous materials can be expected.â23 In March 2012, NTSB recommended that PHMSA require all newly manufactured and existing tank cars used for ethanol and crude oil have enhanced tank head and shell punctureâresistance systems and top fittings protection.24 When NTSB issued this recommendation in 2012, the use of unit trains to move crude oil had been growing dramatically. As the timelines in Table 4-5 indicate, the number of crude oil incidents began to rise as traffic volumes increased. In July 2013, the Lac-MÃ©gantic crash occurredâa crude oil train disaster that killed 47 people. The investigations of that incident again pointed to the failings of the DOT-111 car in resisting damage; how- ever, the 65 mph speed of the runaway train created immense crash forces. Of concern was whether the DOT-111 could resist damage at the lower speeds more commonly associated with derailments. Tables 4-6 and 4-7 show the distribution of train speeds in the crude and ethanol tank cars derailed during 2006 to 2015. Higher speed appears related to the number of cars derailed and damaged, the quantity released, and the occurrence of a fire. Uniquely contributing to this disaster, the high train speed at the time of derailment stemmed from the locomotive engineer failing to adequately apply and test the trainâs hand brakes when stopped and left unattended on a descending grade. However, a number of derailments during 2014 and 2015 in which tank cars sustained considerable damage occurred at relatively low speeds. 23 National Transportation Safety Board, âRailroad Accident Brief: Tiskilwa, Illinois, Octo ber 7, 2011,â Accident Brief NTSB/RAB-13/02 (Washington, D.C.), page 9, accessed Sep- tember 19, 2017, https://www.ntsb.gov/investigations/AccidentReports/Reports/RAB1302.pdf. 24 National Transportation Safety Board, âSafety Recommendation R-12-5 through R-12-8, R-07-4,â March 2, 2012, https://www.ntsb.gov/safety/safety-recs/recletters/R-12-005-008.pdf.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 91 TABLE 4-7 Crude Oil Tank Car Incidents in the United States, 2005â2015 Estimated Speed Before Derailment (mph) 1â10 11â20 21â30 31â40 +40 Number of Derailments 3 3 6 5 4 Cars Incurring Damage or Release 7 7 32 67 24 Release Quantity (gallons) 4,800 88,679 238,051 839,169 502,277 Release Quantity (barrels) 114 2,111 5,668 19,980 11,959 SOURCE: PHMSA, âIncident Statistics.â TABLE 4-6 Ethanol Tank Car Incidents in the United States, 2005â2015 Estimated Speed Before Derailment (mph) 1â10 11â20 21â30 31â40 +40 Number of Derailments 33 7 9 3 6 Cars Incurring Damage or Release 41 7 34 36 53 Release Quantity (gallons) 107,146 82,253 454,299 999,512 976,378 Release Quantity (barrels) 2,551 1,958 10,816 23,798 23,247 NOTE: Speed was reported as zero for five incidents despite all involving moving crashes or derailments. These five were not included in the table under the assumption that speed was reported incorrectly. SOURCE: PHMSA, âIncident Statistics.â In the February 2015 derailment in Mount Carbon, West Virginia, all of the DOT-111 cars were compliant with an upgraded CPC-1232 industry standard, which was adopted in October 2011 for new tank cars transport- ing crude oil and ethanol. The CPC-1232 cars, as discussed in Chapter 3, were equipped with half-height head shields, thicker tank and head steel, fitting protection, and pressure relief devices. According to FRAâs investiga- tion, the train in the Mount Carbon incident was being operated at 33 mph, below the 50-mph maximum authorized speed.25 The cause of the derail- ment was determined to be a broken rail. Twenty-seven of the tank cars carrying crude oil left the track; however, only 6 incurred enough damage to release their contentsâ2 sustaining punctures and 4 sustaining valve or fitting damage. The major consequences in this case occurred after the on- set of a pool fire, which caused thermal tank shell failures on 13 tank cars that had otherwise survived the accident. The thermal tears caused violent 25 U.S. Federal Railroad Administration, âAccident Findings Report for Derailment of CSX Transportation, Inc.âs Unit Crude Oil Train K08014 Transporting Crude Oil for Plains All American, Mount Carbon, West Virginiaâ (Washington, D.C., October 9, 2015), https://www. fra.dot.gov/eLib/Details/L17123.
92 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES fireball eruptions; the first occurring within a half hour of the derailment and the last 10 hours later. Less than a month after the Mount Carbon incident, a crude oil unit train traveling at 23 mph derailed 21 CPC-1232 tank cars in Galena, Illinois. In this case, product released from cars sustain- ing damaged valves and fittings ignited into a pool fire that resulted in five tank car thermal failures. Following a review of these incidents, NTSB concluded that the CPC- 1232 design was deficient for carrying flammable liquids, lacking the needed pressure-relief capacity and thermal insulation to prevent tank failures from overpressurization when exposed to a pool fire. In April 2015, the agency issued urgent recommendations calling for more robust and fire-resistant tank cars for carrying crude oil and ethanol.26 It called for an aggressive schedule of replacing or retrofitting the DOT-111 and the CPC-1232 fleet with cars equipped with thermal insulation and higher-capacity pressure- relief devices. Later in this chapter, a summary is provided of the key provisions of a May 2015 federal rulemaking to address safety issues arising from the in- creased rail movement of ethanol and crude oil by rail.27 Regulators expect the outcome of this rulemaking, which phases out the legacy DOT-111 and CPC-1232 cars and establishes a new DOT-117 design (and retrofit version) with greater crashworthiness and thermal protection, will be fewer tank cars releasing product during derailments and fewer thermal ruptures. The continued potential for severe crashes, however, was highlighted by a June 2016 incident in Mosier, Oregon, in which four CPC-1232 tank cars were breached from a unit train carrying Bakken oil. The derailed cars lost more than 40,000 gallons of product, which ignited and required the evacuation of about 100 people. The cause of the Mosier derailment appears to have been a track failure, the persistence of which has caused NTSB to raise con- cern about the duration of the phase-out schedule, which extends to May 2025. Accordingly, NTSB has called on regulators to schedule intermediate milestones and progress-reporting requirements as a means of encouraging earlier tank car replacements and retrofits. WATERWAYS SAFETY TRENDS Like pipelines, the waterways have long transported crude oil, and to a lesser extent ethanol. Indeed, all imports of crude oil to the United States, 26 National Transportation Safety Board, âSafety Recommendation R-15-14 Through -17,â April 3, 2015, http://www.ntsb.gov/safety/safety-recs/recletters/R-15-014-017.pdf. 27 Pipeline and Hazardous Materials Safety Association, âHazardous Materials: Enhanced Tank Car Standards and Operational Controls for High-Hazard Flammable Trains,â 89 FR 26644, 2015, https://www.gpo.gov/fdsys/pkg/FR-2015-05-08/pdf/2015-10670.pdf.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 93 except for shipments from Canada, are transported in tanker ships. These ships are also used to transport oil from the terminus of the Trans-Alaska Pipeline to refineries in the United States. As discussed in Chapter 3, crude oil and ethanol are transported on the inland and coastal waterways pri- marily by tanker barge. The U.S. Coast Guard (USCG), which is responsible for overseeing and regulating the safe operations of these vessels and movements, has established three major safety measures: the âsheen rule,â designation of a responsible party, and the requirement for double-hulled vessels. Because of the sensitivity of the maritime environment, carriers are required to report any spill or discharge of petroleum and other hazardous substances to the National Response Center (NRC), which is run by USCG. Applying what is commonly known as the âsheen ruleâ (the Discharge of Oil regulation), the agency requires the person in charge of a vessel to report any observed sheen or discoloration on the surface of the water or that causes a sludge of emulsion to be beneath the waterâs surface or on the shoreline, even if the source of the discharge is not known at the time. The report must contain the location of the release, an estimate of the quantity, information about its source and cause, types of material, and data on threats posed and consequences such as injuries. NRC relays the release information in the report to a USCG or U.S Environmental Protection Agency (EPA) On-Scene Coordinator to evaluate the situation and decide whether a federal emer- gency response action is necessary and/or to monitor the cleanup activities of the vessel operator, designated by law as the responsible party, and the local and state governments. As the responsible party, vessel operators must make preparations along their routes to anticipate the possibility of a dis- charge and remediate as necessary. To prevent a recurrence of incidents such as the Exxon Valdez tanker spill and others from the late 1980s, vessels are required to be double-hulled to minimize a loss of containment. These reforms have been credited with not only fundamentally chang- ing the set of safety incentives that maritime carriers and shippers face but also with creating a safety culture and assurance system that is anticipatory so that safety is reasonably assured in the face of unforeseen changes in traffic levels, technology, and operating practices. The safety performance data that follow bear out the effectiveness of these reforms. Figure 4-14 shows the total volume of releases reported by tank barges of any kind of petroleum, including crude oil and refined products, from 1995 to 2015. In most years, the total volume has not exceeded 5,000 bar- rels. In November 2005, a tank barge carrying 119,000 barrels of heavy crude oil from Houston, Texas, to Tampa, Florida, crashed into the remains of a pipeline service platform that was damaged during Hurricane Rita. The
94 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES FIGURE 4-14 Petroleum oil spilled by tank barges in the United States, 1995â2015. SOURCE: American Waterways Operators and U.S. Coast Guard, âU.S. Coast Guard-American Waterways Operators Annual Safety Report: Towing Industry Safety Statistics 1994â2014,â August 3, 2016, 5, http://www.americanwaterways.com/sites/ default/files/USCG%20AWO%20Annual%20Safety%20Report%20August%20 2016.pdf. barge discharged more than 46,000 barrels into the Gulf of Mexico.28 Since this 2005 incident, there have been no comparable petroleum releases from vessels in U.S. waters. When normalized according to barrels of petroleum transported by water, the discharge rate has not exceed 5,000 barrels per billion barrels transported since 2005 (see Figure 4-15). Focusing solely on crude oil discharges, the total annual discharge volume has only exceeded 25 barrels from tankers and barges combined in 2 years since the 2005 Gulf of Mexico spill, as shown in Figure 4-16. Dur- ing 2013, 16 incidents were reported, for a combined discharge of 183.5 barrels. One tank barge incident, on the Mississippi River near Vicksburg, accounted for 171 of these barrels. The tank barge, with a cargo hold con- taining approximately 1,900 barrels of crude oil, struck a railroad bridge, causing the closing of the river during control and cleanup activities.29 28 Entrix, âTank Barge DBL 152 Incident Response Environmental Unit Reportâ (U.S. Coast Guard, January 2010), 2-1, https://casedocuments.darrp.noaa.gov/southeast/dbl152/ pdf/5011%20DBL%20152%20Environmental%20Unit%20Report.pdf. 29 U.S. Coast Guard, âClaim Summary/Determination,â September 22, 2016, 1, https:// www.uscg.mil/npfc/Claims/2016/N13011-0004%20RC%20Paid_Redacted.pdf. 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000 Barrels released Total volume released
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 95 FIGURE 4-16 Crude oil discharges and volumes by tankers and barges in the United States, 2006â2015. SOURCE: USCG, âPort State Information Exchange,â U.S. Coast Guard Maritime Information Exchange, accessed September 19, 2017, https://cgmix.uscg.mil/PSIX/ PSIXSearch.aspx. FIGURE 4-15 Discharge rate for all petroleum oil spilled by tank barges in the United States, 1995â2015. SOURCE: American Waterways Operators and U.S. Coast Guard, âU.S. Coast Guard-American Waterways Operators Annual Safety Report: Towing Industry Safety Statistics 1994â2014,â August 3, 2016, 5, http://www.americanwaterways.com/sites/ default/files/USCG%20AWO%20Annual%20Safety%20Report%20August%20 2016.pdf. 4-16 0 20 40 60 80 100 120 140 160 180 200 0 2 4 6 8 10 12 14 16 18 20 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Incidents Incidents Barrels released Barrels 0 5,000 10,000 15,000 20,000 25,000 30,000 Barrels released per billion barrels Total volume released
96 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES COMPARISON OF MODAL INCIDENT DATA Because pipelines, rail, and water are not used to transport all the com- modities examined in this chapter, a comparison of incident rates and other indicators of safety performance is not always possible. Natural gas and NGL shipments (i.e., unfractionated NGLs) are transported exclusively by pipelines, making cross-modal comparisons impossible. Ethanol shippers rely almost exclusively on railroads and on barges to a much lesser extent. Unlike rail, the latter mode has no record of ethanol discharges, but its transported volumes are relatively low. Because it is transported by all three modes in large volumes, crude oil is the only commodity that is a candidate for cross-modal safety com- parisons. Tables 4-8A, 4-8B, and 4-8C show annual reports of crude oil releases and volumes from pipelines, railroads, and waterways for periods that coincide with the domestic energy revolution. It is important to note that the incident data do not take into account the severity of incidents, which limits cross-modal comparability. Comparability is limited further by differences in incident-reporting criteria. As a practical matter, it is also important to recognize that the modes are not always substitutes for one another in moving crude oil. Comparisons of the safety performance of modes when they are used differently and are not viable substitutes can have limited value for decision making. OTHER SAFETY-RELATED EFFECTS The safety data reviewed above are derived largely from carrier reports of hazardous materials incidents to PHMSA. It is important to keep in mind, however, that the injuries and fatalities that are the direct result of hazard- ous materials releases may not be the only safety-related effects of this traf- fic. In the case of railroads, this impact can be significant. As noted earlier, the Lac-MÃ©gantic crash in Canada killed 47 people in 2013. While no fatalities occurred in the United States as a result of releases from derailed crude oil trains during 2005 to 2015, trains do create other safety impacts on a regular basis, most notably as a result of collisions with pedestrians on rights of way and motor vehicles at grade crossings. To the extent that the increases in crude oil moved by rail have led to more train traffic, there is a likelihood that the increase in trains has led to more of these incidents. Based on FRA data for 2012 to 2015, the combined average risk to trespassers and grade-crossing users (including individuals later determined to have been suicidal) is 1.29 fatalities and 1.58 injuries per million train-
97 T A B L E 4 -8 A C ru de O il In ci de nt s an d R el ea se s pe r T ra ffi c V ol um e fo r Pi pe lin es , 20 05 â2 01 5 Y ea r Sh ip m en ts ( ba rr el s) B ar re ls R el ea se d T ra ffi c (b ill io n ba rr el -m ile s) In ci de nt R at e pe r B ill io n B ar re l- M ile s B ar re ls R el ea se d pe r B ill io n B ar re l- M ile s 20 05 3, 17 0, 89 0, 84 7 10 1, 96 4 1, 56 2 0. 11 65 .2 5 20 06 3, 24 5, 58 3, 76 2 83 ,8 51 1, 58 7 0. 10 52 .8 2 20 07 3, 36 1, 38 8, 08 4 19 ,7 87 1, 44 2 0. 11 13 .7 2 20 08 3, 12 2, 11 0, 36 3 59 ,2 52 1, 59 0 0. 10 37 .2 4 20 09 3, 31 0, 02 3, 74 3 24 ,1 83 1, 49 3 0. 10 16 .1 9 20 10 3, 21 1, 41 6, 07 7 52 ,7 10 1, 63 0 0. 09 32 .3 3 20 11 3, 57 0, 07 9, 71 2 35 ,2 76 1, 67 4 0. 09 21 .0 6 20 12 3, 57 7, 52 1, 74 9 15 ,0 25 1, 78 3 0. 10 8. 42 20 13 3, 69 3, 14 7, 85 2 43 ,0 48 1, 81 0 0. 11 23 .7 7 20 14 3, 97 9, 50 1, 92 3 17 ,5 21 2, 07 3 0. 11 8. 45 20 15 4, 64 7, 14 4, 13 2 20 ,6 68 2, 36 1 0. 11 8. 75 To ta l, 20 05 â2 01 5 38 ,8 88 ,8 08 ,2 44 47 3, 28 5 19 ,0 12 0. 10 24 .8 9 To ta l, 20 05 â2 00 9 16 ,2 09 ,9 96 ,7 99 28 9, 03 7 7, 67 7 0. 10 37 .6 5 To ta l, 20 10 â2 01 5 22 ,6 78 ,8 11 ,4 45 18 4, 24 8 11 ,3 35 0. 10 16 .2 6 SO U R C E : Fe de ra l E ne rg y R eg ul at or y C om m is si on ; PH M SA .
98 T A B L E 4 -8 B C ru de O il In ci de nt s an d R el ea se s pe r T ra ffi c V ol um e fo r R ai lr oa d In ci de nt s, 2 00 5â 20 15 Y ea r Sh ip m en ts ( ba rr el s) B ar re ls R el ea se d T ra ffi c (b ill io n ba rr el -m ile s) In ci de nt R at e pe r B ill io n B ar re l- M ile s B ar re ls R el ea se d pe r B ill io n B ar re l- M ile s 20 05 1, 83 7, 50 0 35 .8 1 0. 60 3. 32 59 .3 8 20 06 1, 79 2, 00 0 1. 67 0. 63 1. 59 2. 65 20 07 1, 56 4, 50 0 0. 01 0. 53 1. 87 0. 02 20 08 5, 26 6, 80 0 1, 93 1. 04 4. 61 1. 73 41 8. 54 20 09 5, 62 5, 90 0 0. 01 4. 9 0. 20 0. 00 2 20 10 19 ,7 61 ,7 00 11 7. 13 23 .8 0. 38 4. 92 20 11 50 ,3 69 ,2 00 93 .5 3 66 .5 0. 50 1. 41 20 12 15 8, 11 7, 40 0 89 .8 4 25 0. 9 0. 35 0. 36 20 13 28 2, 19 2, 40 0 22 ,5 11 .5 0 47 4. 5 0. 25 47 .4 4 20 14 34 1, 34 1, 00 0 1, 37 1. 89 57 1. 8 0. 25 2. 40 20 15 28 3, 64 9, 80 0 14 ,2 44 .8 3 46 8. 0 0. 09 30 .4 4 To ta l, 20 05 â2 01 5 1, 15 1, 51 8, 20 0 40 ,3 97 .2 6 1, 86 7 0. 24 21 .6 4 To ta l, 20 05 â2 00 9 16 ,0 86 ,7 00 1, 96 8. 54 11 .3 2 1. 15 17 3. 90 To ta l, 20 10 â2 01 5 1, 13 5, 43 1, 50 0 38 ,4 28 .7 2 1, 85 5 0. 23 20 .7 1 SO U R C E : A ss oc ia ti on o f A m er ic an R ai lr oa ds ; Fe de ra l R ai lr oa d A dm in is tr at io n an al ys is o f ST B W ay bi ll Sa m pl e da ta ; PH M SA .
99 T A B L E 4 -8 C C ru de O il In ci de nt s an d R el ea se s pe r T ra ffi c V ol um e fo r W at er w ay s, 2 00 6â 20 15 Y ea r Sh ip m en ts ( ba rr el s) B ar re ls R el ea se d T ra ffi c (b ill io n ba rr el -m ile s) In ci de nt R at e pe r B ill io n B ar re l- M ile s B ar re ls R el ea se d pe r B ill io n B ar re l- M ile s 20 06 47 6, 46 5, 83 0 13 5. 71 4 43 1. 60 1 0. 02 8 0. 31 4 20 07 48 6, 32 2, 95 0 2. 85 2 45 2. 68 0. 03 5 0. 00 6 20 08 45 8, 97 2, 15 0 2. 71 6 42 8. 14 8 0. 02 6 0. 00 6 20 09 43 4, 66 1, 08 0 0. 07 38 44 0. 51 7 0. 01 1 0. 00 02 20 10 40 7, 94 6, 66 0 24 .5 10 38 3. 63 6 0. 01 0 0. 06 4 20 11 40 2, 34 1, 10 0 2. 36 4 37 8. 30 8 0. 01 6 0. 00 6 20 12 53 3, 24 5, 82 0 1. 02 9 41 1. 58 2 0. 02 4 0. 00 2 20 13 69 2, 60 4, 85 0 18 2. 48 1 49 8. 88 1 0. 03 2 0. 36 6 20 14 75 2, 00 4, 83 0 1. 52 4 57 6. 90 6 0. 01 6 0. 00 3 20 15 74 7, 26 5, 83 0 10 .5 48 59 2. 23 3 0. 01 7 0. 01 8 To ta l, 20 06 â2 01 5 5, 39 1, 83 1, 10 0 36 3. 80 5 4, 59 4. 49 4 0. 02 2 0. 07 9 To ta l, 20 06 â2 00 9 1, 85 6, 42 2, 01 0 14 1. 35 0 1, 75 2. 95 0 0. 02 5 0. 08 1 To ta l, 20 10 â2 01 5 3, 53 5, 40 9, 09 0 22 2. 45 5 2, 84 1. 55 0 0. 01 9 0. 07 8 SO U R C E : U .S . A rm y C or ps o f E ng in ee rs ; U SC G .
100 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES miles.30 As indicated in Table 4-8B, there were approximately 1,865 billion barrel-miles of crude oil moved by rail from 2008 to 2015. Almost all of this volume represents new railroad traffic, because before 2008 railroads moved only about 0.6 billion barrel-miles per year. If it is assumed this new traffic was moved in 100-car unit trains at 700 barrels per tank car, the result would be about 26.6 million additional train-miles during 2008 to 2015, an increase of about 0.5 percent in total freight train-miles. Based on average risks of collisions with pedestrians and motor vehicles cited above, this added traffic would be expected to lead to an additional 34 fatalities and 42 injuries on railroad rights of way and grade crossings over that same period. These estimates are best viewed as rough approximations of the mag- nitude of the safety impact. It is possible that some of this crude oil traffic crowded out other rail traffic, such as movements of grain or coal, so that train-miles were not proportionally increased. In addition, some of the oil traffic could have simply led to longer trains rather than more trains. It is also conceivable that the risk of train collisions with pedestrians and vehicles does not change proportionately with the number of trains in the manner implied by the use of average annual risk factors. Nevertheless, the magnitudes of the estimates suggest that even with these uncertainties, the safety impacts are likely to be large and deserving of consideration by policy makers when interpreting and comparing the modal safety data in Tables 4-8A, 4-8B, and 4-8C. MAJOR RAIL SAFETY INITIATIVES IN RESPONSE TO INCIDENTS The rail, pipeline, and water modes have long been the subject of federal regulations to ensure that hazardous materials shipments are transported safely. Indeed, most public policies that are intended to ensure the safe transportation of crude oil, ethanol, and natural gas have been in place for decades, preceding the post-2005 upsurge in their traffic volumes. The safety data examined above suggest that these baseline measures have been robust. The growth in hazardous materials traffic after 2005 has not been accompanied by comparable growth in the number of hazardous materials incidents or in the number of severe incidents. Nevertheless, the safety statistics show how the location and modal distribution of incidents have changed in ways that reflect new shipping origins, destinations, and modal use patterns. These changes have prompted new policy responses, 30 U.S. Federal Railroad Administration, âOffice of Safety Analysis,â Federal Railroad Administration, February 23, 2017, http://safetydata.fra.dot.gov/officeofsafety/default.aspx. Data were obtained through queries of Casualty Summary Tables, Suicide Casualties by State/ Railroad, and Operational Data Tables.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 101 particularly to ensure the safety of flammable liquids shipments made in tank car unit trains, which as recently as 2005 were not a consideration of rail safety regulators.31 In the case of rail, FRA and PHMSA have increased safety inspec- tions and issued new regulations, emergency orders, and safety advisories targeted specifically to crude oil and ethanol traffic. They have sponsored research to gain a better understanding of the risks associated with trans- porting these commodities and reached out to state and local agencies to help meet their emergency response planning, information, and training needs. Some of these important safety initiatives are summarized next. High-Hazard Flammable Trains Rule In May 2015, FRA and PHMSA jointly issued a new set of regulations in a rulemaking titled Enhanced Tank Car Standards and Operational Con- trols for High-Hazard Flammable Trains (HHFT).32 The rule is intended to be comprehensive by including provisions intended to prevent tank car derailments, limit the severity of incidents when they do occur, and assist state and local agencies in planning and preparing a safer and more effec- tive emergency response to incidents. Some of the major provisions of this rule are summarized next. The rule contains two provisions intended to reduce the likelihood of derailments involving unit trains moving tanks cars containing crude oil or ethanol. First, railroads are required to apply a 27-factor analysis when selecting routes for these trains to travel. The 27 factors are the same ones used for routing trains containing cars loaded with poison gases (toxic-by- inhalation hazards, TIHs). These factors pertain to conditions such as the routeâs traffic density, maintenance, grade, and curvature that can affect the potential for a derailment. Second, the rule requires that train speeds be restricted to 50 mph in all areas, which is the same as the limit for TIH trains. While the main purpose of the speed limit is to reduce the severity of incidents, lower train speeds are also viewed as having the potential to prevent some incidents such as over-speed derailments. Although not part of the HHFT rulemaking, in 2015 FRA also launched the Crude Oil Route Track Examination (CORTEx) program to further its goal to prevent incidents. This program concentrates increased track inspec- tions on crude oil routes by a team of inspectors. Each round of inspections 31 U.S. Federal Railroad Administration, âSafe Placement of Train Cars: Report to the Senate Committee on Commerce, Science, and Transportation and the House Committee on Transportation and Infrastructureâ (Washington, D.C.: U.S. Department of Transportation, June 2005), https://www.fra.dot.gov/eLib/Details/L03467. 32 Hazardous Materials: Enhanced Tank Car Standards and Operational Controls for High- Hazard Flammable Trains, p. 26645.
102 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES lasts about 2 weeks and includes an audit of the railroadâs track inspection records. Afterward, regional inspectors are instructed to reinspect items that had been identified by the CORTEx team. In addition to reducing the occurrence of incidents, several provisions in the HHFT rule are intended to reduce the severity of incidents when they do occur. A main purpose of the aforementioned 50-mph speed restriction is that it limits the kinetic energy in a crash. Indeed, the rule requires that maximum speeds be reduced from 50 mph to 40 mph in urban areas when the train contains tank cars that do not meet the ruleâs upgraded design standards for crash resistance. Perhaps the most significant provision in the HHFT rule that is aimed at reducing crash incident severity is an upgraded design specification for tank cars used in crude oil and ethanol service.33 As discussed in Chapter 3, the two main designs used for transporting these commodities have been the DOT-111 and the CPC-1232 specifications. The former is a general service nonpressure design, while the latter is a version of the same design enhanced with a stronger tank shell, added head and fittings protection, and a pressure-relief device. The 2014 HHFT rule created the new DOT- 117 standard that contains several enhancements to these legacy standards that are intended to increase resistance to tank punctures (e.g., full-height head shields, thicker tank shells), reduce overpressurization from exposure to heat from fires (e.g., thermal jackets, larger pressure-relief devices), and minimize crash-related damage to top and bottom fittings (e.g., added protection for top fittings and bottom outlet valves designed to prevent unintended releases or have the handle removed before moving). The rule also includes a new specification (i.e., DOT-117R) for retrofitting DOT-111 and CPC-1232 cars to bring them closer to the DOT-117. These design changes were informed in large part by analyses of industry data on more than 47,000 tank cars damaged in incidents since the 1960s.34 However, these data included few records of tank cars in crude oil or ethanol unit train derailments. A summary of the DOT-117/DOT-117R standards as they compare to the DOT-111 and the CPC-1232 is provided in Table 4-9. In creating these new tank car specifications, the HHFT rule contains a design- and commodity-specific phase-out schedule for DOT-111 and 33 Pipeline and Hazardous Materials Safety Administration, âHazardous Materials: Enhanced Tank Car Standards and Operational Controls for High-Hazard Flammable Trains Notice of Proposed Rulemaking,â 79 FR 45016 (2014), https://www.regulations. gov/contentStreamer?documentId=PHMSA-2012-0082-0180&disposition=attachment& contentType=pdf. 34 Christopher Barkan and Todd Treichel, âRail Transportation of Crude Oil and Ethanolâ (Committee for a Study of the Domestic Transportation of Petroleum, Natural Gas, and Ethanol, Washington, D.C., May 12, 2016), 6, http://www.trb.org/PolicyStudies/ CommitteeMeetings2.aspx.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 103 TABLE 4-9 Changes to Tank Car Design in 2015 HHFT Rule Car Features Legacy DOT-111 (non-jacketed or jacketed) â¢ 7/16â³ tank shell, no jacket or â¢ 7/16â³ tank shell, jacket CPC-1232 (non-jacketed or jacketed) â¢ 1/2â³ tank shell, no jacket, half-height head shield, top fittings protection or â¢ 7/16â³ tank shell, jacket, full-height head shield, top fittings protection DOT-117 â¢ 9/16â³ tank shell, jacket, full-height head shield, top fittings protection, thermal protection, top and bottom fittings protection â¢ Retrofitted DOT-117R permits a tank shell thickness of 7/16â CRC-1232 cars that is intended to account for differences in the relative risks of each design when transporting crude oil and ethanol. Crude oil may not be transported in any DOT-111 car starting in early 2018 or in any CPC-1232 car that lacks thermal protection starting in early 2020. Ethanol may not be transported in DOT-111 cars or in CPC-1232 cars that lack thermal protection in 2023. By mid-2025, all CPC-1232 cars are prohibited from transporting either crude oil or ethanol. It merits noting that Transport Canada announced an expedited phase-out of all DOT-111 (TC-111 in Canada) tank cars used in crude oil service by moving up the deadline to November 1, 2016.35 To supplement the intended safety benefits of the DOT-117 tank car specifications, the HHFT rule also includes a provision to reduce the severity of derailments through a requirement that HHFTs be equipped with one of the following advanced braking systems: distributed power or a two-way end-of-train device. The rule additionally requires that all HHFT unit trains (HHFUTs; each with 70 or more tank cars loaded with flammable liquids) be equipped with electronically controlled pneumatic (ECP) brakes by 2021.36 The ECP brake requirement, which was advocated by NTSB,37 seeks to limit the consequences of derailments by reducing the number of derailed cars, kinetic energy, and the potential for tanks to suffer damage and re- 35 Progressive Railroading, âCanada to Speed up Phaseout of DOT-111s for Crude Oil Use,â Progressive Railroading, http://www.progressiverailroading.com/safety/news/ Canada-to-speed-up-phaseout-of-DOT-111s-for-crude-oil-use--48943. 36 HHFUTs used to transport ethanol are required to implement ECP braking by 2023. 37 National Transportation Safety Board, âTrain Braking Simulation Studyâ (National Trans- portation Safety Board, July 20, 2015), http://dms.ntsb.gov/public/55500-55999/55926/577439. pdf.
104 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES lease commodity. Seeking a stronger justification for the rule, Congress has required the Secretary of Transportation to conduct addi tional testing and analysis of ECP braking systems when applied to HHFUTs and to decide by the end of 2017 whether to retain the requirement. As inherently chaotic events, train derailments pose a particularly formidable challenge to test and analyze through means such as simulation and modeling. Asked to review the U.S. Department of Transportationâs (U.S. DOTâs) efforts, a sepa- rate National Academies of Sciences, Engineering, and Medicine committee found that the results from the modeling and simulations lacked sufficient validation to enable confident estimation of the emergency performance of ECP braking systems compared with that of pneumatic braking systems aug- mented with distributed power or end-of-train devices.38 The ECP braking committee was unable to make a conclusive statement about the emergency performance of ECP brakes relative to other braking systems on the basis of the results of testing and analysis provided by U.S. DOT. Also, to address concerns that crude oil from formations such as Bakken may be more volatile than oil from more conventional meansâand thus more likely to ignite when released in an incidentâthe HHFT rule requires that shippers develop and document a sampling and testing program for the identification and characterization of unrefined petroleum oil shipments. Previously, shippers relied on general crude oil safety data sheets not neces- sarily representative of the crude oil characteristics from certain sources, such as the Bakken formation. Additionally, PHMSA has collaborated with the U.S. Department of Energy to sponsor research by Sandia National Lab- oratories to identify gaps in crude oil characterization and sampling methods for further research.39 An initial finding of the Sandia research is that many factors are relevant to the flammability of crude oil in a tank car derailment, and thus concern about the characteristics of tight oil, particularly from the Bakken formation, may be too narrowly focused.40 Track Safety Standards FRA issued new track inspection and maintenance regulations in 2014 that apply to hazardous materials and passenger routes. A prominent feature 38 Transportation Research Board, âA Review of the Department of Transportation Plan for Analyzing and Testing Electronically Controlled Pneumatic Brakes; Letter Report (Phase 2)â (Washington, D.C.: Transportation Research Board, September 29, 2017), https://www.nap. edu/catalog/24903. 39 U.S. Department of Energy, âCrude Oil Characteristics Research,â Energy.Gov, July 9, 2015, http://energy.gov/fe/articles/crude-oil-characteristics-research. 40 David Lord et al., âLiterature Survey of Crude Oil Properties Relevant to Handling and Fire Safety in Transportâ (Sandia National Laboratories [SNL-NM], Albuquerque, NM, March 2015), 83, http://www.osti.gov/scitech/biblio/1177758.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 105 of the rule focuses on establishing a performance-based standard for the maximum allowable frequency of track service failures.41 The regulation also sets intervals between track inspections, criteria for remedial action for defective rails, and training qualifications for rail flaw inspectors.42 In addition, FRA administers a program to promote implementation of innovative inspection technologies. This programâin use by most Class I railroadsâallows railroads to petition to pilot the use of innovative means of nonstop continuous rail testing for inspections, such as by gage restraint measurement systems and automated track geometry vehicles.43 Automa- tion, for instance, has the potential to enable more frequent inspections than required by current regulation, which would have the side benefit of enabling the prioritized reallocation of railroad inspection personnel to the routes frequently traversed by unit trains moving large quantities of hazard- ous energy liquids. In briefings to the committee, one railroad reported that the use of automated track inspection vehicles had increased its inspection capability as measured in miles of track from 91,000 miles in 2014 to ap- proaching 240,000 miles by the end of 2016. Train Securement The inadequate application of hand brakes on a crude oil unit train stopped overnight en route to Lac-MÃ©gantic allowed the equipment to roll down the track about 7 miles before derailing in town. As one of the causal factors in- volved in the derailment of the runaway train, in August 2015, FRA issued a rule, Securement of Unattended Equipment, to regulate the securement of unattended equipment to mitigate this risk unique among other recent crude oil derailments.44 This rule requires hand brakes or other mechanical securement devices to be sufficient to hold the equipment, as well as that the controlling locomotive hauling any hazardous materials be locked when stopped and unattended outside of a yard. 41 U.S. Federal Railroad Administration, âTrack Safety Standards: Improving Rail Integrity,â 49 CFR 213 (2014), https://www.fra.dot.gov/Elib/Document/3546. The performance standard requires 0.09 or fewer service failures per year per mile of track for hazardous materials routes or 0.08 or fewer service failures per year per mile of track for routes with hazardous materials and passengers sharing a route (p. 4234). 42 Ibid., p. 4248. The rulemakingâs discussion about setting inspection intervals does not provide an obvious technical basis for setting 30 million gross tons as the standard as opposed to the previous standard of 40 million gross tons or some alternative figure. 43 Karl Alexy, Director, Office of Safety Analysis, FRA, in discussion with project staff, Oc- tober 6, 2016. The program allows railroads to pilot nonstop continuous rail-testing processes in accordance with 49 CFR Parts 211 and 213. 44 U.S. Federal Railroad Administration, âSecurement of Unattended Equipment,â 49 CFR 232 (2015), https://www.regulations.gov/contentStreamer?documentId=FRA-2014-0032-0012& disposition=attachment&contentType=pdf.
106 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES MEASURES TO AID EMERGENCY PREPAREDNESS AND RESPONSE The Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA) and the Oil Pollution Act of 1990 (OPA 1990) form the basis for regulations issued by EPA, FRA, PHMSA, and USCG governing emergency preparedness and response to spills of oil and other hazardous materials from transportation. Among other things, the regulations define agency and industry roles in national and regional response teams and estab- lish the unified command structure across agencies and the responsible party. An important element of this contingency planning is the National Response System, which is activated by the report of an incident to USCGâs NRC. Figure 4-17 lists the various entities that coordinate within this sys- tem, the relevant mode, and their role in preparedness or response. The details of these national planning and preparedness programs are too complex to present here. It merits noting, however, that special regulatory provisions deal with the concerns associated with oil discharges into water. USCG, for instance, requires waterways operators to develop response plans, and PHMSA requires pipeline operators to do the same. FIGURE 4-17 Primary domestic governmental, private sector, and non governmental organizations involved in the National Response System as it relates to long-distance crude oil, natural gas, and ethanol shipments. NOTE: Involved with pipelines = (p); railroads = (r); waterways = (w); preparedness = *; response = â . 4-17 Governmental â¢ U.S. Environmental Protection Agency (EPA) *â (p)(r) â¢ U.S. Coast Guard (USCG) *â (p)(w) â¢ Pipeline and Hazardous Materials Safety Administration (PHMSA) *â (p)(r) â¢ Federal Railroad Administration (FRA) *â (r) â¢ Federal Emergency Management Agency (FEMA) â â¢ National Oceanic and Atmospheric Administration (NOAA) *â (p)(r)(w) â¢ State Emergency Response Commissions (SERCs) *(p)(r)(w) â¢ Local Emergency Planning Committees (LEPCs) *â (p)(r)(w) â¢ State, tribal, and local responders (SEMAs; e.g., fire and rescue and police departments) *â (p)(r)(w) Private Sector / Trade Associations â¢ Association of American Railroads (AAR) *(r) â¢ American Waterways Operators (AWO) *(w) â¢ American Oil Pipe Lines (AOPL) *(p) â¢ American Petroleum Institute (API) *(p)(r) â¢ Interstate Natural Gas Association of America (INGAA) *(p) â¢ Renewable Fuels Association (RFA) *(r)(w) â¢ Oil Spill Response Organizations (various) *â (p)(r)(w) Nongovernmental Organizations â¢ CHEMTREC (CHEMical TRansportation Emergency Center) â (p)(r)(w) â¢ International Association of Fire Chiefs (IAFC) *â (p)(r)(w) â¢ National Association of SARA Title III Program Officials (NASTTPO) *(p)(r)(w) â¢ National Association of State Fire Marshals (NASFM) *(p) â¢ National Fire Protection Association (NFPA) *â (p)(r)(w) â¢ TRANSCAER (TRANSportation Community Awareness and Emergency Response) *(r)(p)
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 107 The planning requirement entails documenting that adequate resources and expertise as necessary are available to mitigate and recover from a worst- case oil discharge into water. In the case of waterways, these plans usually include oil spill trajectory modeling. These oil spill dispersion models en- able response organizations to more effectively anticipate where to marshal skimmers and booms to protect the environment and recover crude oil.45 While railroads are not subject to similar spill planning regulation, PHMSA proposed a new rule in July 2016, as summarized in Box 4-1, that would introduce such requirements for this mode. This proposed HHFT rule included two additional provisions intended to improve the preparedness and procedures for emergency responses to rail incidents involving crude oil. First, railroads would be required to provide state and tribal agencies with a point of contact for HHFT shipment infor- mation, similar to the provision for TIH shipments. Second, the rule would make permanent the May 2014 emergency order that required railroads to inform State Emergency Planning Commissions about any single train carry- ing 1 million or more gallons of Bakken crude oil. The notification must in- clude information on routes, shipment frequencies, and points of contact.46 The information-sharing policy in the May 2014 emergency order was apparently developed without consultation with state and tribal authorities responsible for emergency response preparation. In a letter to PHMSA, the National Association of SARA Title III Program Officials (NASTTPO) expressed concern about the inadequacies of information sharing by rail carriers pertinent to HHFT traffic. The letter maintained that state offi- cials were not being provided timely commodity flow information, contact information for railroad hazardous materials managers, and information about railroad emergency preparedness plans and response capabilities.47 To the extent that railroads shared information, NASTTPO claimed that state officials were often hampered in their dissemination of the information to local emergency responders because of uncertainty about the sensitivity of the information and the receipt of redacted documents.48 In some cases, railroads shared information with state fusion centers, which are law 45 The use of oil spill trajectory modeling was prompted by the 1976 spill by the Argo Merchant off the coast of Nantucket, Massachusetts. 46 In July 2016, PHMSA proposed a rule that would expand the Bakken crude oil notifica- tion requirement to cover all HHFTs. 47 National Association of SARA Title III Program Officials, âHazardous Materials: Oil Spill Response Plans for High-Hazard Flammable Trains,â September 20, 2014, 1, https:// www. regulations.gov/contentStreamer?documentId=PHMSA-2014-0105-0206&attachment Number=1&disposition=attachment&contentType=pdf. 48 Government Accountability Office, âEmergency Responders Receive Support, but DOT Could Improve Oversight of Information Sharing,â 35, http://www.gao.gov/assets/690/681123. pdf.
108 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES Box 4-1 Oil Spill Response Plans for High-Hazard Flammable Trains Notice of Proposed Rulemaking Under current regulations, nearly all shipments of oil by rail need a basic oil spill response plan (OSRP), an internal document identifying the means of and personnel and equipment available for response to a discharge. This proposal by PHMSA and FRA would go beyond that for HHFT and accomplish two main purposes: Comprehensive OSRPs: Railroads would develop these plans for FRA approval consistent with federal guidance on spill response and planning to ensure the capacity to respond to and remove a worst-case discharge, accounting for at least 300,000 gallons of petroleum oil. Although the OPA 1990 framework omits ethanol, the proposed rule would include denatured alcohol fuel in a concentration of as little as 10 percent petroleum oil. The proposed rule would bring rail ship- ments into alignment with the comprehensive oil spill response plan requirements established under OPA 1990. Petroleum Oil Testing: It would allow an American Society for Testing and Materials (ASTM) standard for determining a productâs initial boiling point to be used to test petroleum oil that more accurately measures more volatile compo- nents in crude oil (i.e., the âlight endsâ). Although this provision would not incor- porate the American Petroleum Instituteâs recommended practice for testing of petroleum oil for railroad tank cars, the addition of the ASTM initial boiling point test would bring regulation and industry practice closer. enforce ment bodies that do not have strong ties to state and local authori- ties responsible for emergency response planning and preparations. In most cases, local and state agencies are first on the scene for a pipe- line rupture or train derailment. PHMSA provides grants to state and local authorities to develop and test emergency response plans and to train fire- fighters and police on response methods and practices. PHMSA has created two new programs to deliver training to responders to crude oil and ethanol transportation incidents. The agencyâs Assistance for Local Emergency Re- sponse Training grant programs target training of emergency responders, with a focus on rural and volunteer responders, whose increased participa- tion in these training activities is a priority for policy makers and industry. The Transportation Rail Incident Preparedness and Response program pro- vides train-the-trainer resources and distributes information online to the emergency response community on crude oil properties and recommended response practices such as hazard assessment, risk evaluation, and choice of protective clothing and equipment.
SAFETY PERFORMANCE OF LONG-DISTANCE TRANSPORTATION 109 Shippers and carriers of crude oil and ethanol have also taken steps to improve the training of responders. A notable example is the railroad- and shipper-sponsored TRANSCAER program,49 which is dedicated to improv- ing understanding of and response capabilities for hazardous materials transportation incidents. Between 2014 and 2015, TRANSCAER provided training for more than 100,000 responders.50 As rail shipments of ethanol and crude oil increased, TRANSCAER increased its focus on these flam- mable liquids. For example, the program partnered with the American Petroleum Institute to create a crude-by-rail safety course that has been presented at national, state, and local hazardous materials and emergency response conferences. To further assist with this response, PHMSA publishes the Emergency Response Guidebook (including a version for mobile devices),51 which provides initial response guidance on the hazards involved, actions to be taken, and public protection options. To ensure that first responders have shipment hazard information at the scene of incidents, PHMSA requires that hazardous shipments be accompanied by shipping papers that describe the material and its hazards and provide the shipperâs emergency contact information. These papers usually reference the American Chemistry Coun- cilâs CHEMTREC,52 a 24-hour hotline for emergency responders to obtain information about chemicals and hazardous materials. Most railroads have standardized their shipping papers to report the train position of each car carrying hazardous material, and they typically designate a single point of contact for routing decisions so that state and regional fusion centers and other officials are aware of the routes used to transport hazardous materials. The Association of American Railroads (AAR) assisted in the development of the mobile app AskRail, which pro- vides details on a trainâs consist and links the listed cargo to response infor- mation from the Emergency Response Guidebook to enable faster decision making on-scene (see Figure 4-18).53 AAR released AskRail in late 2014. 49 TRANSportation Community Awareness and Emergency Response. 50 Frank Reiner, âTRANSCAER: Assisting Communities to Prepare for and Respond to Pos- sible Hazardous Material Transportation Incidentsâ (Committee for a Study of the Domestic Transportation of Petroleum, Natural Gas, and Ethanol, Washington, D.C., February 5, 2016), 6, http://www.trb.org/PolicyStudies/CommitteeMeetings2.aspx. In 2015, TRANSCAER pro- grams trained 52,474 responders in addition to the nearly 55,000 trained in 2014. (https:// www.transcaer.com/docs/general/TRANSCAER-FactSheet-5.pdf). 51 Pipeline and Hazardous Materials Safety Administration, â2016 Emergency Response Guidebookâ (U.S. Department of Transportation, 2016), http://www.phmsa.dot.gov/staticfiles/ PHMSA/DownloadableFiles/Files/Hazmat/ERG2016.pdf. 52 CHEMical TRansportation Emergency Center. 53 Bob Fronczak, âRail Transport of Petroleum, Natural Gas and Ethanolâ (Committee for a Study of the Domestic Transportation of Petroleum, Natural Gas, and Ethanol, Washington, D.C., February 5, 2016), 26.
110 SAFELY TRANSPORTING HAZARDOUS LIQUIDS AND GASES Despite these efforts, first responders encounter difficulty acquiring the necessary training to manage incidents involving tank car unit trains. The time commitment needed for training and exercises can limit participation, especially for volunteer firefighters. According to a survey of 23 counties conducted by the U.S. Government Accountability Office (GAO), many first responders, including a majority of volunteer firefighters in rural areas, cited the need to take unpaid leave from work to attend training sessions.54 Other reasons given for not taking advantage of training opportunities include the difficulty that both volunteer and professional units face in filling the positions of responders who are away at training and obtaining permission for leave from regular duties and traveling outside the local community. While most of the 23 counties surveyed by GAO reported that their first responders had the basic competencies needed to identify hazard- ous cargoes and notify appropriate parties in the event of an incident, 17 reported that less than 60 percent of their personnel had more advanced levels of response expertise and training.55 54 Government Accountability Office, âEmergency Responders Receive Support, but DOT Could Improve Oversight of Information Sharing,â 17. 55 Ibid., p. 15. FIGURE 4-18 AARâs AskRail app screenshots for train consist and tank car details, hazardous materials identification, and response guidance resources (left to right). 4-18