Marathon Pipe Line LLC-9/3/08

Page 1Official PDFDOT US Department of Transportation PHMSA Pipelines and Hazardous Materials Safety Administration OPS Office of Pipeline Safety Central Region Principal Investigator James Bunn/Karen Butler Region Director Ivan Huntoon (Prior RD) Date of Report 1/14/2015 Subject Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Operator, Location, & Consequences Date of Failure 09/03/2008 Commodity Released Hazardous Liquid (Refined Products, Diesel) City/County & State Ashmore / Coles, Illinois OpID & Operator Name 32147 Marathon Pipe Line LLC Unit # & Unit Name 3703 Wabash System SMART Activity # 122652 Milepost / Location MP 41.7 / 100 yards downstream of Ashmore Pump Station Type of Failure Material Failure, Rupture - Crack in a dent, Deformation from poor construction practice Fatalities 0 Injuries 0 Description of Area impacted Total Costs $2,936,068 Rural farmland, non-High Consequence Area (HCA)#

Page 2Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 Executive Summary At approximately 9:43 a.m. CDT1 on September 3, 2008, a 12-inch diameter refined products pipeline operated by Marathon Pipe Line Company LLC (Marathon) ruptured downstream of the Ashmore pumping station located in Coles County, Illinois. Marathon’s Findlay Operations Center received suction and discharge pressure deviation alarms from the Ashmore Pump Station at 9:43. At 9:44 a.m. the Operations Center received a CPM leak alarm, and the Wabash Pipeline System (12 inch) was shut down. Marathon personnel were dispatched to locate the accident site. The National Response Center (NRC) was contacted at 12:58 p.m. (report number 882609). The rupture was confirmed at 5:12 p.m. to be located approximately 100 yards downstream of the Ashmore Pump Station on the Robinson to Champaign section of the pipeline known as the East Leg. The pressure at the time of the failure was 1,158 psig with a maximum operating pressure (MOP) of 1,211 psig. A second NRC report was filed (report number #882650) at 6.08 p.m., which increased the reported spill volume and provided additional details on the location. No fatalities, injuries, fire, explosion, or evacuations resulted from the rupture, and it did not occur in a High Consequence Area (HCA). Surface water was impacted (1 barrel). Approximately 770 barrels of low sulfur diesel fuel were released as a result of the rupture, and an estimated 2,902 cubic yards of soil was contaminated. There were no service interruptions or supply impacts as a result of the rupture. The total cost estimated by the operator as a result of the release was $2,936,068. The pipeline was constructed in 1958 and was composed of 12-inch nominal diameter, 0.250-inch wall- thickness, low frequency electric resistance welded (LF ERW), X46 pipe made by Republic Steel. Prior to the failure, the East Leg had experienced six hydrostatic tests in addition to that at original construction (starting in 1964 and ending in 2004). These six additional hydrostatic tests resulted in 26 ruptures and six leaks. At the rupture location, a deformation was present from original construction. Coating damage and approximately 5 percent through-wall external corrosion was identified at the center of a deformation. After the metallurgical report was completed, the operator engaged two additional subject matter experts to review the failure and metallurgical analysis. Three subject matter experts arrived at a “consensus that the failure was caused by a crack located in a deformation. The deformation was centered adjacent to the longitudinal ERW seam. Part of the deformation intersected the ERW seam. The deformation likely occurred prior to the original commissioning of the pipeline. The shear crack formed on the outside diameter of the pipe due to a partial re-rounding of the deformation anomaly. Early hydrotests may have imparted compressive stresses in the area of the crack that retarded subsequent fatigue crack growth during operation. The fracture surfaces between the beach marks do not look like those typical of high-cycle fatigue from operating pressure cycles. The apparent crack advancement is indicative of low-cycle fatigue. The crack grew by tearing in one or more of the multiple hydrostatic tests between 1958 and 2004. The crack eventually grew to a large enough size that operational pressure continued to advance the crack to the point of failure.”2 System Details Marathon Pipe Line Company LLC operates over 5,000 miles of hazardous liquid pipelines in 15 States. 1 All times reference Central Daylight Time unless otherwise specifically noted. 2 Information provided by Marathon during the follow-up from recommendations identified during the internal failure investigation. Page 2 of 7#

Page 3Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 The Wabash 12-inch Products Pipeline System originates at two points, Robinson, Illinois and Wood River, Illinois. The Wabash System terminates in Hammond, Indiana. The West Leg is 129.9 miles long and is located between Wood River, Illinois, and Champaign, Illinois. The North Leg is 128.4 miles long with a 109.8 mile section located between Champaign, Illinois, and Griffith, Indiana, and another section that is 18.6 miles long located between Griffith, Indiana, and Hammond, Indiana. The East Leg is 85.1 miles long and located between Champaign, Illinois, and Robinson, Illinois. This rupture occurred on the East Leg of the system. The East Leg has pumping stations at Robinson, Martinsville, and Ashmore. Product from Robinson, Illinois, can go into tankage at Champaign or can bypass Campaign and continue on to Hammond, Indiana. The East Leg was constructed with 12-inch nominal diameter pipe with a wall thickness of 0.250 inches. The pipe was manufactured in 1958 by Republic Steel Corporation and is LF ERW, API 5LX, grade X-46 line pipe. The pipeline was joined using the shielded metal arc welding process, and the welds were inspected by radiography. The pipe is coated with coal tar enamel, and the pipeline has been under cathodic protection since 1958. The pipeline was hydrostatically tested originally in 1958. A hydrostatic test program was developed to manage the threat of manufacturing defects associated with the longitudinal weld seam, and additional tests were performed in 1964, 1974, 1987, 1995, 1999, and 2004. The MOP at the time of the failure was established as 1,211 psig and based on 80 percent of test pressure. The pipeline was inspected with a High Resolution MFL and Geometry tool in 2002. There were no anomalies identified at the failure location during this ILI inspection. The pipeline was inspected by PHMSA in 2006. Events Leading up to the Failure Just prior to the incident, the East Leg was delivering an 80,000 bbl batch of ultra low sulfur diesel to the Champaign Terminal. The Hammond Terminal Headgate Valve was opened in preparation for a turn into the North Leg out of the Champaign Terminal. Following the completion of the Champaign batch, the East Leg was turned out of the Champaign Terminal northward toward the Hammond Terminal by opening the Champaign Bypass Valve. At this time, Robinson, Martinsville, and Ashmore pump stations had one unit running at each location. After turning northward out of the Champaign Terminal, units were started at Champaign, Buckley, and St. Anne pump stations (Buckley and St. Anne are on the North Leg). The hydrostatic test pressure in 2004 was 1,562 psig with a spike test at 1,727 psig. The high pressure shutdown on discharge pressure was recorded as set at 1,220 psig for Ashmore station. The pressure at the time of failure on the discharge of Ashmore pump station was approximately 1,158 psig. Spike tests were used in the hydrostatic test program starting in 1995 and were 30 minutes long at approximately 95 percent SMYS. As the years progressed, hydrostatic tests were performed in smaller sections of pipeline on the East Leg in an effort to help locate releases or leaks. Each test starting in 1964 had more than one failure in the Robinson to Champaign section. The six additional hydrotests starting in 1964 resulted in 26 ruptures and six leaks and are summarized as follows: In 1964 – Two ruptures occurred. This test was conducted with fuel oil but later water was used to complete the test. In 1974 – Six ruptures and one leak occurred. In 1987 – Three ruptures occurred. Page 3 of 7#

Page 4Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 In 1995 – Four ruptures occurred during the spike test. No failures occurred during the 8-hour test. In 1999 – Six ruptures occurred on the spike test, and one leak occurred during the 8-hour test. The causes of these failures were three hook cracks and four lack of fusion defects. In 2004 – Five ruptures occurred on the spike test (three lack of fusion and two hook cracks) and four leaks during the 8-hour test (associated with the longitudinal seams and repaired with Type B sleeves). A backpressure control valve did not exist at Champaign for the Robinson to Champaign segment. Backpressure control methods (check valves or control valves) were also not available at several other locations on the North Leg (Ashkum Junction did not have a check valve; there was no backpressure control valve was available at Hammond). Emergency Response The Operations Center received a CPM leak alarm at 9:44 a.m. on September 3, 2008, and the Operations Center initiated shut down activities for the entire Wabash System within 47 seconds according to report details. Field personnel were notified and dispatched to investigate. Information regarding the incident was reported by Marathon to the National Response Center (NRC) on September 3, 2008, at 12:58 p.m. (Report No. 882609) even though detailed location information was not available and identified two creeks as possible locations for product (Deer Creek and Greasy Creek). At approximately 5:12 p.m. on September 3, 2008, field personnel reported that a rupture was discovered approximately 100 yards downstream of the Ashmore Pump Station. Additional time was required to locate the leak as it was in a cornfield and the release could not be easily identified from the air. Field personnel walked the right-of-way. Marathon filed a second NRC report on September 3, 2008, at 6:08 p.m. (Report No. 882650) providing additional information on the location of the release and an estimated spill volume of 900 barrels. This NRC report also stated that only soil was impacted. Marathon personnel immediately took steps to contain the product in the area where the release occurred. A series of interceptor trenches were dug in the cornfield at the release location, booms were deployed, and the product was contained by 8:00 p.m. The remote control valves upstream and downstream of the leak were closed. The length of the segment that was isolated was 79,200 feet (15 miles). The only emergency responders active at the response scene were Marathon employees. No fatalities, injuries, fire, explosion, or evacuations resulted from the rupture, and it did not occur in a High Consequence Area (HCA). No significant road closures were reported. Surface water (1 barrel) was impacted and approximately 2,902 cubic yards of soil were contaminated. There were no service interruptions or supply impacts as a result of the rupture. Summary of Return to Service After the location of the release was confirmed and the area made safe, the failure site was excavated and the origin of the rupture was identified. The section of pipe that contained the origin was removed and protected, then shipped to a metallurgical laboratory for examination. A new section of pre-tested pipe (51 feet long) was installed in the line where the rupture occurred and the line was returned to service at 3:39 p.m. on September 4, 2008, with a 20 percent pressure reduction. Page 4 of 7#

Page 5Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 Clean-up efforts continued at the site and soil continued to be remediated. Investigation Details At approximately 9:43 a.m. on September 3, 2008, Marathon Pipe Line Company LLC’s Findlay Operations Center received suction and discharge pressure deviation alarms from its Ashmore Pump Station located in Coles County, Illinois. At 9:44 a.m. the Operations Center received a CPM leak alarm and the Wabash Pipeline System was shut down and Marathon personnel were dispatched to locate the incident site. Control Center personnel responded to the events within 36 seconds according to report data, and within 47 seconds, flow was stopped on the pipeline. Marathon personnel discovered the rupture location at 5:12 p.m. as the pipeline right-of-way was walked downstream of the Ashmore Pump Station. The rupture occurred on the pipeline right-of-way approximately 100 yards downstream of the pump station at Survey Station Number 2207+54, MP 41.7. The rupture was reported as 70 inches long with an opening approximately 3.0 inches in width at the 5:30 circumferential position. The deformation measured approximately 0.150 inches deep (1.2 percent of pipe OD at the deepest location). Coating damage and approximately 5 percent through-wall external corrosion was present at the center of the deformation. No indications of internal corrosion, stress corrosion cracking, or other outside force damage were identified during metallurgical analysis. Residual stresses when the pipeline was cut were reported to be negligible. The second NRC volume estimating 900 barrels would later be determined to be an estimated 770 barrels of low-sulfur diesel. Initially after the failure, the pressure setpoints were lowered on all three legs of the Wabash system. The pressure de-rates remained in place while the operator ran ILI tools and completed the associated rehabilitation work. After the metallurgical analysis was completed, Marathon also completed an internal failure investigation, which resulted in three recommendations: 1) utilize ILI technology to identify other potential dents in the vicinity of the ERW seam and remediate any deformations as appropriate, 2) collaborate with Scheduling and Hydraulics to perform an evaluation of potential changes to operating procedures and/or engineering controls that would reduce the frequency and magnitude of pressure cycles, and 3) employ additional independent resources to obtain a second opinion on the metallurgical findings. 3 Beginning in October of 2008 (10/26), deformation ILI assessments were completed on the North, East and West Legs. The results of these caliper runs were overlaid with circumferential MFL data in order to identify any deformation at the long seam that might be similar to the release location. As a result, a total of 66 repairs were performed, 35 of which were on this same section (East Leg of the pipeline). An additional hydrostatic test was completed in November of 2008 on a portion of the North Leg (Champaign to Griffith). The operator hired two additional Subject Matter Experts (SMEs) to review the metallurgical analysis. The three SMEs agreed that “the failure was caused by a crack located in a deformation. The 3 Identified in the Marathon investigation summary report. Page 5 of 7#

Page 6Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 deformation was centered adjacent to the longitudinal ERW seam. Part of the deformation intersected the ERW seam. The deformation likely occurred prior to the original commissioning of the pipeline. The shear crack formed on the outside diameter of the pipe due to a partial re-rounding of the deformation anomaly. Early hydrotests may have imparted compressive stresses in the area of the crack that retarded subsequent fatigue crack growth during operation. The fracture surfaces between the beach marks do not look like those typical of high-cycle fatigue from operating pressure cycles. The apparent crack advancement is indicative of low-cycle fatigue. The crack grew by tearing in one or more of the multiple hydrostatic tests between 1958 and 2004. The crack eventually grew to a large enough size that operational pressure continued to advance the crack to the point of failure.”4 A pressure trending review did identify higher pressures at the Ashmore pump station versus other stations in the East line section just prior to the failure, which may indicate that the changes in operations that occurred at Champaign caused pressure to back into the Ashmore pump station operations even though the maximum MOP was not exceeded. Marathon conducted a detailed review of potential operational changes that could be made through scheduling similar batches of like product together in an effort to minimize starts and stops (reducing potential pressure cycles) and also determined that several control enhancements could assist in the reduction of amplitude regarding pressure cycles. Suggested control revisions included replacement of the Griffith backpressure control valve, a new check valve downstream of the Ashkum Junction, and new valves added at Hammond and Champaign on the East Leg. Pressure setpoints were also reviewed for possible changes. The operator purchased the property that this rupture occurred on, both for clean-up efforts with continued soil remediation and for future ease of access. Findings & Contributing Factors The metallurgical analysis determined that the fracture originated in a deformation in the pipe wall near the longitudinal seam. The deformation occurred during the original construction of the pipeline. A shear crack formed on the outside diameter of the pipe due to re-rounding of the deformation, which likely occurred during the initial pressure test. Three SMEs agreed that “early hydrotests may have imparted compressive stresses in the area of the crack that retarded subsequent fatigue crack growth during operation. The fracture surfaces between the beach marks did not look like those typical of high-cycle fatigue from operating pressure cycles. The apparent crack advancement is indicative of low-cycle fatigue. The crack grew by tearing in one or more of the multiple hydrostatic tests between 1958 and 2004. The crack eventually grew to a large enough size that operational pressure continued to advance the crack to the point of failure.”5 The lack of adequate back pressure controls may have been a contributing factor, and Marathon’s internal failure investigation provided recommendations to minimize this potential and reduce pressure cycles. As a result of additional ILI assessments, a total of 66 repairs were performed where deformation and longitudinal seams may have overlapped, 35 of which were on this same segment. 4 Marathon follow-up recommendations identified during the internal failure investigation. 5 Information provided by Marathon during the follow-up from recommendations identified during the internal failure investigation. Page 6 of 7#

Page 7Failure Investigation Report – Marathon Pipe Line LLC – Material Failure, Rupture Failure Date 09/03/2008 An additional hydrostatic test was completed in 2008. The six hydrostatic tests occurring in 1964 through 2004 resulted in 26 ruptures and six leaks. The operator’s management program involving hydrostatic tests removed defects from the pipeline, which also could have resulted in additional safety and environmental consequences. Appendices A Maps and Photographs B NRC Reports C Operator’s Accident Report D Metallurgical Analysis Page 7 of 7#

Page 8Appendix A - Maps and Photographs Wabash 9-3-2008 (b) (/)(F)| Legend Point Of Release 145 290 Location: llfdyrs981\gis_datalwo Page 1 of 7#

Page 9Appendix A - Maps and Photographs 1500m L KENDALL MUSCATINE LA PORTE: ISLAND OCK BUREAU LA SALLE WILL JAKE JIRTER MERCER PUTNAM GRUNGY -STARKE ES STARK MARSHALE KANKAKEE ST- ANNE P JASPER INES KNOX PULASKI WARREN WOODFORD LIYINGSTON PEORtA ASHKUM TRODS WHITE BENTON CARROI MCDONOUG ILLINOIS CLEAN * BUCKLEY CHAMPATON FORD WARREN MASON: LOGAN DEWITT VERMICCION CLINT SCHUTLER GROWN CeSs MENARD I FOUNTAIN PIAT • CHAMPAIGN MACON MORGEN SANDAMON 0OU5LAS 104 PARKE (SCOTT WEST LEG "MACON RE • ASHORE EDGAr PUTNAM I (b) (7X(F) COLES EAST LEG VIGO CUMBERLAND CLAY CLARK MARTINSVILLE OWEN LINCOLN ULIVAN MOWAd HAM JASPEC CRAWF CHARLES LAWRE OPERATOR MAP Page 2 of 7#

Page 10Appendix A - Maps and Photographs (b) (7)(F) Ashmore Incident Resnonse Man O#

Page 11Appendix A - Maps and Photographs Failed Pipeline Joint Page 4 of 7#

Page 12Appendix A - Maps and Photographs Close-up photo showing rupture site Page 5 of 7#

Page 13Appendix A - Maps and Photographs Field, north side of Ashmore station (provided by operator) Internal view of removed pipe (provided by operator) Page 6 of 7#

Page 14Appendix A - Maps and Photographs Image of tie-in (provided by operator) Page 7 of 7#

Page 15Appendix B - NRC Reports TeleDetail Page 1 of 2 VPHMSA Administration Materials Safely Pipeline & Hazardous HMIS->INCIDENTS->TELEPHONICS (Version 4.0.0 PROD) Rules of Behavior Home Logout Ment. NRC Number: Call Date: 882609 [Retum to Search] 09/03/2008 Call Time: 13:58:00 Caller Information Address Company Name MARATHON PIPELINE LAURA PERRON City 539 S. MAIN ST. Country FINDLAY Phone 1 USA Conndental Organization Type PRIVATE 2172511885 ZIp Phone 2 45840 Ores No O No Response Is caller the spller? © Yes O No O No Response First Name Address Company Name LAURA Discharger Information MARATHON PIPELINE Last Name PErroN Country City FINDLAY 539 S. MAIN ST. Phone 1 USA zip State OH Organization Type PRIVATE 2172511885 Phone 2 45840 Nearest City State Spill Information Location OAKLAND County Zip Code CoLES DTG Type Spll Date 09/03/2008(mmidayyy) Incident Type ALL <- Select DTG Type -> v Spll Time 09 45 00 (24hh mm ss) Descriotion Reported Incident Type PIPELINE THE CALLER IS REPORTING A POTENTIAL RELEASE OF ULTRA LOW SULFUR DIESEL FUEL FROM PRESSURE. LOCATE ANY POSSIBLE LEAKS. AN UNDERGROUND 12 INCH FLOW LINE. THE LINE HAS BEEN SHUT DOWN AND EMPLOYEES ARE CURRENTLY ATTEMPTING IO THE POTENTIAL WAS CAUSED BY A DROP IN PIPELINE laterals Invoive Chris Code Total Qty. Water Ory. Medium Type Addltional Medium Information -Select Medium [POTENTIAL RELEASE TO THE SOIL, DEER CREEK AND GREASY CREEK Evacuations Injures Damages Federal Agency Notted O Yes O No Unknown Other Agency Noted O Yes O No Unknown State Agency Noted O Yes No Unknown Remedal Actions Page 1 of 4#

Page 16Appendix B - NRC Reports TeleDetail Page 2 of 2 EMPLOYEES ARE IN THE PROCESS OF SEARCHING THE LINE FOR ANY SINGS OF A RELEASE NONE Adattonal Inte Degrees Latttude Lenatude Minutes seconds Quadrant Distance from City Degrees Seconds Range Section Township Direction quadrant Milepost • Rescinded Commenta (max 250 characters) ex Previous 1.1 of 1 Page 2 of 4#

Page 17Appendix B - NRC Reports TeleDetail Page 1 of 2 VPHMSA Administration Materials Safely Pipeline & Hazardous HMIS->INCIDENTS->TELEPHONICS (Version 4.0.0 PROD) Rules of Behavior Home Logout Ment. NRC Number: Call Date: 882650 [Retum to Search] 09/03/2008 Call Time: 19:08:59 Caller Information Address Company Name MARATHON PIPELINE LAURA PERRON City 539 S. MAIN ST. Country FINDLAY Phone 1 USA Conndental Organization Type PRIVATE 2172511885 ZIp Phone 2 45840 Ores No O No Response Is caller the spller? © Yes O No O No Response First Name Address Company Name LAURA Discharger Information MARATHON PIPELINE Last Name PERRON City 539 S. MAIN ST. Phone 1 Country USA zip State OH Organization Type PRIVATE 2172511885 Phone 2 State Spill Information Nearest Cly Location ASHMOORE Zip Code County 10151 NORTH COUNTY ROAD 2170 E Spll Date Incident Type DTG Type 09/03/2008(mm/dayyy) ALL <- Select DTG Type -> v Spll Time [15 00 00 (24hh mm ss) Reported Incident Type PIPELINE UNKNOWN CAUSES. CALLER IS REPORTING A DISCHARGE OF ULTRA LOW SULFER DIESEL FROM A PIPELINE DUE TO Matertals Involved Material / Chris Name DIL DIESEL loos "Chris Code 1900 BARRELS Total Qty Water ory. Adational Medium Information Medium Type <- Select Medium SOIL Evacuations Injures Fatalites Damages Ores No Unknown Federal Agency Noted O yes O No ® Unknown Other Agency Notted Ores ONo Unknown State Agency Noted O Yes O No Unknown Bemedal Actions Page 3 of 4#

Page 18Appendix B - NRC Reports TeleDetail Page 2 of 2 WAS SHUT IN. CAUSE OF DISCHARGE IS UNDER INVESTIGATION. BEGINNING CLEANUP OPERATIONS. LINE Adattonal Inte CALLER HAD NO ADDITIONAL INFORMATION Degrees Latitude Lenatude Minutes seconds Quadrant Distance from City Degrees Seconds section Direction Quadrant Range Township Milepost • Rescinded Commenta (max 250 characters) ex Previous 1.1 of 1 Page 4 of 4#

Page 19Appendix C - Operator's Accident Report U.S Department of Transportation Pipeline and Hazardous Materials Safety Administration ACCIDENT REPORT – HAZARDOUS LIQUID PIPELINE SYSTEMS Report format corresponds to Form PHMSA F 7000-1 (01-2001) Original Report Date October 03, 2008 No. 20080292 - 14467 PART A – GENERAL INFORMATION N Original Report Y Supplemental Report Y Final Report Last Revision Date 01/14/2015 1. Operator Name and Address a. Operator's 5-digit Identification Number 32147 b. If Operator does not own the pipeline, enter Owner's OPS 5-digit Identification Number (if known) c. Name of Operator MARATHON PIPE LINE LLC d. Operator street address 539 SOUTH MAIN STREET e. Operator address City FINDLAY County or Parish HANCOCK State OH Zip code 45840 2. Time and date of the accident Hour 09:44 Date of the accident 09/03/2008 3. Location of accident a. Latitude 39.523801 Longitude -88.070495 b. City ASHMORE County or Parish COLES c. State IL Zip Code 61912 d. Mile Post/Valve Station Survey Station No 2207+54 4. Telephone Report NRC Report Number 882609 Date 09/03/2008 5. Losses (Estimated) Public/Community Losses reimbursed by operator Public/private property damage $ 118,587 Cost of emergency response phase $ 2,543,148 Cost of environmental remediation $ 38,875 Other Costs $ 0 Describe Operator Losses Value of product lost $ 81,620 Value of operator property damage $ 99,802 Other Costs $ 54,036 Describe LAND PURCHASE SUPPORT Total Costs $ 2,936,068 6. Commodity Spilled Commodity spilled (yes/no) Y a. Name of commodity spilled ULTRA LOW SULFUR DIESEL b. Classification of commodity spilled GASOLINE, DIESEL, FUEL OIL OR OTHER PETROLEUM PRODUCT WHICH IS A LIQUID AT AMBIENT CONDITIONS c. Estimated amount of commodity involved Unit of Measure BARRELS Amount Spilled 770.00 Amount Recovered 770.00 CAUSES FOR SMALL SPILLS NO DATA PART B – PREPARER AND AUTHORIZED SIGNATURE 1 Page 1 of 6#

Page 20Appendix C - Operator's Accident Report Preparer's Name JENNIFER RADER Area Code and Telephone Number 4194214631 Preparer's E-mail Address JERADER@MARATHONOIL.COM Area Code and Facsimile Number 4194213917 PART C – ORIGIN OF THE ACCIDENT 1. Additional location information a. Line segment name or ID ASHMORE-CHAMPAIGN 12 b. Accident on Federal Land other than Outer Continental Shelf NO c. Is pipeline Interstate Y Offshore N d. Area Block # State Outer Continental Shelf N 2. Location of system involved Operator's Property NO Pipeline Right of Way Y High Consequence Area (HCA) N Describe HCA 3. Part of system involved in accident ONSHORE PIPELINE, INCLUDING VALVE SITES Other (specify) If failure occurred on Pipeline, complete items a-g a. Leak or Rupture RUPTURE Type of Leak - Puncture, diameter (inches) Type of Rupture LONGITUDINAL-TEAR/CRACK - Tear/Crack, length (inches) 3 - Propagation Length, total, both sides (feet) 6 Other (specify) b. Type of block valve used for isolation immediate section Upstream Manual NO Automatic NO Remote Control YES Check Valve NO Downstream Manual NO Automatic NO Remote Control YES Check Valve NO c. Length of segment isolated (ft) 79200 d. Distance between valves (ft) 316800 e. Is segment configured for internal inspection tools? YES f. Had there been an in-line inspection device run at the point of failure? YES g. If Yes, type of device run High Resolution Magnetic Flux tool YES Year run 2002 Low Resolution Magnetic Flux tool NO Year run UT tool NO Year run Geometry tool YES Year run 2002 Caliper tool NO Year run Crack tool NO Year run Hard Spot tool NO Year run Other tool NO Year run 4. Failure occurred on PIPE SEAM Other (specify) Year the component that failed was installed 1958 5. Maximum operating pressure (MOP) 2 Page 2 of 6#

Page 21Appendix C - Operator's Accident Report a. Estimated pressure at point and time of accident (PSIG) 1158 b. MOP at time of accident (PSIG) 1211.00 c. Did an over pressurization occur relating to the accident? N PART D – MATERIAL SPECIFICATION 1. Nominal pipe size (NPS) (inches) 12.75 2. Wall thickness (inches) .25 3. Specification API 5L SMYS 46000 4. Seam type ERW LOW FREQUENCY 5. Valve type N/A 6. Manufactured by REPUBLIC STEEL in year 1958 PART E - ENVIRONMENT 1. Area of accident UNDER GROUND Other (specify) 2. Depth of cover (inches) 4 PART F - CONSEQUENCES 1. Consequences Fatalities Injuries a. Number of operator employees 0 0 Contractor employees working for operator 0 0 General public 0 0 Totals 0 0 b. Was pipeline/segment shutdown due to leak? Y If Yes, how long? Days 1 Hours 5 Minutes 22 c. Product ignited Gas did not Ignite d. Explosion NO EXPLOSION e. Evacuation (general public only) N Number of people Reason for Evacuation f. Elapsed time until area was made safe Hours Minutes 2. Environmental Impact a. Wildlife Impact Fish/aquatic N Birds N Terrestrial N b. Soil Contamination Y If Yes, estimated number of cubic yards 2902 c. Long term impact assessment performed N d. Anticipated remediation Y If Yes, check all that apply Surface Water N Groundwater N Soil Y Vegetation Y Wildlife N e. Water Contamination Y Amount in water (barrels) 1 Ocean/Seawater N Surface Y Groundwater N Drinking water N Drinking water source PART G – LEAK DETECTION INFORMATION 1. Computer based leak detection capability in place? Y 2. Was the release initially detected by? CPM/SCADA-BASED SYSTEM WITH LEAK DETECTION 3 Page 3 of 6#

Page 22Appendix C - Operator's Accident Report Other (specify) 3. Estimated leak duration Days 0 Hours 1 PART H – APPARENT CAUSE H1 – CORROSION 1. External Corrosion 2. Internal Corrosion Complete items a-e where applicable a. Pipe Coating b. Visual Examination Other (specify) c. Cause of Corrosion Other (specify) d. Was corroded part of pipeline considered to be under cathodic protection prior to discovering accident? Year Protection Started e. Was pipe previously damaged in the area of corrosion? Estimated time prior to accident Years Months H2 – NATURAL FORCES 3. Earth Movement Description Other (specify) 4. Lightning 5. Heavy Rains/Floods Description Other (specify) 6. Temperature Description Other (specify) 7. High Winds H3 – EXCAVATION DAMAGE 8. Operator Excavation Damage (including their contractors / Not Third Party) 9. Third Party a. Excavator group b. Type Other (specify) c. Excavation was d. Excavation was ongoing activity (Month or longer) If Yes, Date of last contact e. Did operator get prior notification of excavation activity? If Yes; Date received null Notification received from f. Was pipeline marked? i. Temporary markings ii. Permanent markings iii. Marks were iv. Were marks made within required time? H4 – OTHER OUTSIDE FORCE DAMAGE 10. Fire/Explosion as primary cause of failure Fire/Explosion cause 11. Car, truck or other vehicle not relating to excavation activity damaging pipe 12. Rupture of Previously Damaged Pipe 13. Vandalism H5 – MATERIAL AND/OR WELD FAILURES Material 14. Body of Pipe 4 Page 4 of 6#

Page 23Appendix C - Operator's Accident Report Description Other (specify) 15. Component Description Other (specify) 16. Joint Description Other (specify) Weld 17. Butt Description Other (specify) 18. Fillet Description Other (specify) 19. Pipe Seam Yes Description LF ERV Other (specify) Complete a-g if you indicate any cause in part H5 a. Type of failure Construction Defect N Description POOR WORKMANSHIP Material Defect Y b. Was failure due to pipe damage sustained in transportation to the construction or fabrication site? N c. Was part which leaked pressure tested before accident occurred? Y d. Date of test Year 2004 Month 10 Day 12 e. Test medium WATER Other (specify) f. Time held at test pressure (hr) 8 g. Estimated test pressure at point of incident (PSIG) 1514 H6 – EQUIPMENT 20. Malfunction of Control/Relief Equipment Description Other (specify) 21. Threads Stripped, Broken Pipe Coupling Description Other (specify) 22. Seal Failure Description Other (specify) H7 – INCORRECT OPERATION 23. Incorrect Operation a. Type Other (specify) b. Number of employees involved who failed a post-accident test Drug test Alcohol test H8 - OTHER 24. Miscellaneous Describe 25. Unknown Describe PART I – NARRATIVE DESCRIPTION OF FACTORS CONTRIBUTING TO THE EVENT THE FINDLAY OPERATIONS CENTER RECEIVED A SUCTION AND DISCHARGE PRESSURE DEVIATION ALARMS AT ASHMORE STATION. A CPM LEAK ALARM FOLLOWED SHORTLY AFTER AND THE WABASH SYSTEM WAS IMMEDIATELY SHUTDOWN. SHUTDOWN OF THE SYSTEM WAS 5 Page 5 of 6#

Page 24Appendix C - Operator's Accident Report INITIATED 36 SECONDS AFTER CPM ALARM AND FLOW WAS STOPPED 11 SECONDS LATER. THE WABASH EXPERIENCED A FAILURE 100 YARDS DOWNSTREAM OF ASHMORE STATION. THE LINE WAS TURNED FROM CHAMPAIGN TO HAMMOND IMMEDIATELY PRIOR TO FAILURE. ANALYSIS OF THE PIPE REVEALED THAT THE FAILURE MECHANISM WAS A DEFORMATION ANOMALY THAT WAS CENTERED ADJACENT TO THE LONGITUDINAL ERW SEAM WITH A SHEAR CRACK THAT FORMED ON THE OUTSIDE DIAMTER OF THE PIPE DUE TO RE-ROUNDING OF THE DEFORMATION. A LONGITUDINAL SEAM SPLIT DID OCCUR, AND WAS THE PRIMARY CONTRIBUTOR TO THE VOLUME RELEASED; HOWEVER, THE SEAM SPLIT WAS SECONDARY TO THE DEFORMATION ANOMALY CENTERED ADJACENT TO THE LONGITUDINAL ERW SEAM. A REVISED FINAL REPORT IS BEING SUBMITTED THIS 23RD DAY OF MARCH, 2010 CHANGING THE CAUSE FROM MISCELLANEOUS TO RUPTURE, LONGITUDINAL SEAM SPLIT. THE DEFORMATION IS BELIEVED TO HAVE OCCURRED DURING THE ORIGINAL COMMISSIONING HYDROSTATIC PRESSURE TEST. THEREFORE, CONSTRUCTION DEFECT WAS SELECTED THOUGH NEITHER OPTION AVAILABLE IS TRULY REPRESENTATIVE OF THE SITUATION. POOR WORKMANSHIP WAS SELECTED AS THE SPECIFIC CONSTRUCTION DEFECT SINCE THE DEFORMATION OCCURRED DURING THE ORIGINAL CONSTRUCTION AND THE CRACKING PROPAGATION OCCURRED FROM REROUNDING DURING HYDROSTATIC PRESSURE TESTS. BECAUSE THIS REPORTING FORM DOES NOT ALLOW FOR REPORTING OF MINUTES IN QUESTION G3, THE DATA HAS BEEN INPUT AS LEAK DURATION OF 1 HOUR. HOWEVER, AS NOTED IN THIS NARRATIVE, THE LINE WAS SHUTDOWN UPON THE ALARM SIGNALING A FAILURE. WATER CONTAMINATION IS ESTIMATED AT LESS THAN A BARREL, BUT DUE TO THE CONSTRAINTS OF THE FORM, IT IS LOGGED AS 1 BARREL. RUPTURE PROPOGATION LENGTH WAS DETERMINED TO BE 70 INCHES, WHICH IS 5.8 FEET. DUE TO CONSTRAINTS ON THE FORM, PROPOGATION LENGTH WAS ROUNDED UP TO 6 FEET. MARATHON REPORTED A SECOND NRC REPORT (#882650) UPDATING SPILL VOLUME. 6 Page 6 of 6#

Page 25Appendix D Metallurgical Analysis This document is on file at PHMSA#