Environ. Sci. Technol. 2007, 41, 7242-7251
Use of Chemical Fingerprinting to Establish the Presence of Spilled Crude Oil in a Residential Area Following Hurricane Katrina, St. Bernard Parish, Louisiana S C O T T A . S T O U T , * ,† B O L I U , † GLENN C. MILLNER,‡ DYRON HAMLIN,‡ AND EDWARD HEALEY† NewFields Environmental Forensics Practice, LLC, 100 Ledgewood Place, Suite 300, Rockland, Massachusetts 02370, and Center for Toxicology and Environmental Health, LLC, 5120 North Shore Drive, North Little Rock, Arkansas 72118
Hurricane Katrina’s storm surge displaced and damaged a 250 000 barrel storage tank causing a Nigerian crude oil blend (API 36.4°) to be released and dispersed into the adjacent evacuated residential area by the retreating floodwaters. The subsequent environmental assessment involved sampling and chemical fingerprinting of nearly 15 000 wipe and soil samples collected both inside and outside of buildings to determine which properties were impacted by the spilled crude oil. Tier 1 qualitative analysis of gas chromatograms and Tier 2 quantitative (revised Nordtesttype) and qualitative (ASTM D5739-type) analysis of petroleum biomarkers revealed the extent of crude oil contaminationsas well as the widespread occurrence of hydrocarbons derived from (i) lubricating, hydraulic, and transmission oils, most likely from vehicles in the flooded area, and (ii) allochthonous natural organic matter (NOM) from the surrounding bayous. Conventional oil spill fingerprinting protocols and two-component mixing models (crude oil/lube oil and crude oil/NOM) were used to confirm the presence of the spilled crude oilseven when mixed at low concentrations with other hydrocarbon sourcessas a means to develop and govern a settlement and remedial program with the affected property owners.
Introduction The source and impact of waterborne oil spills intentionally or accidentally discharged into the environment are frequently investigated using chemical fingerprinting (1). The success of fingerprinting investigations stems from the chemical specificity of different crude oils and refined petroleum products (2), which affords the opportunity to correlate a spilled oil with its potential source(s) and with potentially impacted environmental matrices, usually sediments and biota, sometimes years after the spill event (3-7). The events following the landfall of Hurricane Katrina on August 29, 2005 included dozens of oil spills of varying magnitudes and circumstances throughout coastal Louisiana. One of these spills was remarkable in that a significant * Corresponding author phone: (781)681-5040; e-mail: sstout@ newfields.com. † NewFields Environmental Forensics Practice. ‡ Center for Toxicology and Environmental Health. 7242
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
waterborne oil spill occurred in a residential areastwo circumstances that normally are mutually exclusive. Assessing the impact of this spill posed many challenges, not the least of which was (1) establishing the extent of the spilled crude oil and (2) distinguishing the crude oil from other non-crude oil hydrocarbon sources in the area, both of which are the objectives of this study. Murphy Oil, Inc. owns and operates the 100 000 barrel/ day Meraux Refinery in St. Bernard Parish, Louisiana. The City of Chalmette (population pre-Katrina over 32 000) lies just west of the refinery, about 7 miles east of New Orleans. Crude oil storage at the refinery occurs primarily in three 250 000 barrel (bbl) and two 450 000 bbl above ground storage tanks (ASTs). Hurricane Katrina’s approximately 17 ft storm surge dislodged one of the 250 000 bbl AST from its concrete foundation, moved it about 35 feet, whereupon it ruptured upon contacting a small concrete sump. The tank contained approximately 40 000 bbl of a Nigerian Bonnie Light crude oil blend (API 34.6°), and gauging subsequently indicated that approximately 25 000 bbl was released. Some crude oil was retained by the AST dikes, but some was transported to the west and into the adjacent residential area of Chalmette along with the receding floodwater. Floating oil impacted structures in the area, leaving a visible “bathtub ring” of oil coating many upright surfaces. In addition, large quantities of particulate matter carried from nearby bayous and lakes within the floodwater facilitated oil-particle interactions leading to sedimentation of some oil onto the ground surface. Assessing the extent of impact of the crude oil in the adjacent community was a critical component in the management of the cleanup, Murphy’s settlement program with affected residents, and class action litigation (now settled). Our assessment began on September 9, 2005 and quickly revealed the presence of hydrocarbons and organic materials unassociated with the spilled crude oil (e.g., allochthonous organic plant debris and peat carried from nearby bayous and deposited both inside and outside of buildings and oil leaked from more than 1000 abandoned vehicles and countless other oil sources, e.g., individual garages, auto shops, etc. in the area). Thus, the objectives of the chemical fingerprinting component of the assessment, described herein, were to develop a reliable means of distinguishing the spilled crude oil from these other potential sources of hydrocarbons and to recognize the presence of even small amounts crude oil when mixed with other hydrocarbon sources.
Materials and Methods Samples. In the 9 months following Hurricane Katrina (September 2005 to May 2006), an assessment of the surrounding community, which included 14 712 samples, was conducted following a work plan approved by the Louisiana Department of Environmental Quality (LDEQ) and the U.S. EPA. Two types of potentially contaminated (biased) samples were collected (viz., 8696 soils and 6016 wipes), both having been collected from the exterior and interior of structures in the study area. Soil samples targeted surficial flood deposits, rather than pre-existing grade soils, and consisted of compositing at least three exterior and three interior samples from a given property, biased to include those that appeared impacted with oil whenever present. The wipes were collected using cotton swabs that were rubbed over darkened “bathtub ring” stains on structures in the area. All sample locations were recorded using photography and a hand-held global positioning system (GPS), and sample 10.1021/es070909i CCC: $37.00
2007 American Chemical Society Published on Web 10/04/2007
FIGURE 1. Chemical fingerprints of samples containing Murphy crude oil. (A-D) Tier 1 GC/FID chromatograms of samples containing Murphy crude oil in variable states of weathering and mixing and (E-H) corresponding Tier 2 GC/MS partial ion chromatograms (m/z 191). #: n-Alkane carbon number; UCM: unresolved complex mixture; NHOP: 17r(H),21β(H)-norhopane; HOP: 17r(H),21β(H)-hopane; OL: oleanane; MOR: 17β(H),21r(H)-moretane; and IS: internal standards. Concentrations listed are in milligrams per kilogram of oil. data were managed using handheld personal digital assistants. The candidate source crude oil samples collected included the unweathered crude oil that remained inside the failed AST (n ) 1) and weathered crude oils and exterior tank wipes (n ) 7) collected from the diked AST area. Because of the pervasive occurrence of allochthonous organic detritus in the study area, peat samples (n ) 21) from the bayous north and east of Chalmette also were collected. All samples were placed in pre-cleaned glass jars and stored and shipped cold (
