http://aapgbull.geoscienceworld.org/content/95/11/1907.full?

HITS=10&hits=10&field_nam e=fulltext&field_value=sediment&searchid=1&FIRSTINDEX=0&resourcetype=HWCIT& src=gr&titlefield=lemhwcomptitle%252Clemextcomptitle&abstractfield=lemhwcompabstr act%252Clemextcompabstract&fulltextfield=lemcontent%3fx=0&gswsubscriber=true&fie ld_name=fulltext&field_value=sediment&src=gr&y=0

Surface sediment hydrocarbons as indicators of subsurface hydrocarbons: Field calibration of existing and new surface geochemistry methods in the Marco Polo area, Gulf of Mexico
1. Michael A. Abrams1 and 2. Dahdah Nicola F.2 + Author Affiliations 1. Exploration and Production Technology Apache Corporation, Houston, Texas; (Michael.Abrams@apachecorp.com) 2. 2Energy and Geoscience Institute University of Utah Salt Lake City, Utah; (ndahdah@egi.utah.edu) Next Section
1

Abstract
Multiple methods are currently used to collect, prepare, extract, and analyze near-surface migrated hydrocarbons from marine sediments to evaluate subsurface petroleum generation and entrapment. Few have been rigorously tested to evaluate their effectiveness. A Gulf of Mexico field calibration survey over the Marco Polo field was undertaken as part of an industry-funded research project to better understand previously published and unpublished seabed geochemical results and determine which gas and liquid hydrocarbon extraction methods best characterize migrated hydrocarbons in near-surface sediments. The Marco Polo calibration data set demonstrates the importance of targeted coring and sampling depth. To improve the detection of seabed migrated thermogenic hydrocarbons, core samples should be collected along major migration pathways (cross-stratal leakage features) identified by conventional deep seismic and high-resolution sea floor imaging. Not all targeted cores hit the designated feature, and thus, collecting replicates along key migration features is

critical. Collecting sediment samples below the near-surface transition zone known as the zone of maximum disturbance is also important to avoid possible alteration effects and interference by recent organic matter. Geochemical analysis should include a full range of hydrocarbon types: light hydrocarbon gases (C1C5), gasoline range (C5C10+), and high-molecular-weight (HMW) hydrocarbons (C15+). The interstitial sediment gas data should be plotted on a total hydrocarbon gas ( C1C5) versus wet gas fraction ( C2C5/ C1C5) chart to identify background, fractionated, and anomalous populations. Compound-specific isotopic analysis on selected anomalous samples is critical to correctly identify migrated subsurface gases from near-surface generated microbial gases. Microdesorption bound gases did not provide gas compositions or compound-specific isotope ratios similar to the Marco Polo reservoir gases, and thus, the bound gas extraction is not recommended. A gasoline range analysis provides a new range of hydrocarbons rarely examined in surface geochemical studies that assist in identifying thermogenic hydrocarbons. Extraction gas chromatography and total scanning fluorescence (TSF) maximum fluorescence intensity provided information on the presence of thermogenic HMW hydrocarbons but did not work as well with the low-level microseepage samples. The TSF fluorogram signature was similar for both seep and regional reference (background) samples and did not help to identify migrated thermogenic hydrocarbons. The Marco Polo calibration study provides a framework to better understand how best to collect (targeted deep cores) and extract migrated hydrocarbons from near-surface marine sediments and to evaluate the results. Previous SectionNext Section

INTRODUCTION
Multiple methods are currently applied to collect, prepare, extract, and analyze near-surface migrated hydrocarbons from marine sediments to evaluate subsurface petroleum generation (Horvitz, 1985; Brooks and Carey, 1986; Abrams et al., 2001; Abrams, 2005; Logan et al., 2009; and Abrams and Dahdah, 2010). Many of the sediment hydrocarbon extraction procedures currently used by the industry are based on sampling and laboratory protocols initially designed for well cuttings and have not always been rigorously tested to evaluate their effectiveness with unconsolidated marine sediments. An extensive series of literature and data reviews, laboratory tests, and field studies have been conducted as part of an industry-funded Surface Geochemistry Calibration (SGC) research project conducted by the University of Utah's Energy and Geoscience Institute. The multiphase SGC research project was organized to better understand previous seabed geochemical results both published and unpublished and determine which sediment hydrocarbon (gas and liquid) extraction methods best characterize migrated nearsurface sediment hydrocarbons (Abrams, 2002). The SGC Gulf of Mexico Marco Polo field calibration survey was designed to field test specific analytical procedures examined in the laboratory studies as well as several emerging and existing technologies. The Gulf of Mexico Marco Polo field was chosen because it is an area of known petroleum leakage based on previous seismic and geochemical surveys (Chaouche et al., 2004;

Dembicki and Samuels, 2007), reservoir geochemical data available (Abrams and Dembicki, 2006), and high-quality shallow seismic imaging data acquired (Dembicki and Samuels, 2007). The near-surface sediment geochemical methods examined in the Marco Polo field calibration survey can be subdivided into three categories: light hydrocarbon gas (C1C5), gasoline-range hydrocarbons (C5C10+), and high-molecular-weight (HMW) hydrocarbons (C15+). The light hydrocarbons were examined using two different sediment gas extraction methods: interstitial (conventional can headspace and modified headspace method called disrupter analysis) and adsorbed-bound (microdesorption analysis). The gasoline-range hydrocarbons were examined by two different analytical procedures, disrupter headspace solid-phase microextraction (HSPME) and Gore Module. The HMW hydrocarbons were examined using solvent extraction followed by whole-extract gas chromatography (GC) and total scanning fluorescence (TSF). The purpose of this article is to provide an overview of the Marco Polo field SGC survey results for different seepage types and to better understand how the different methods characterize the seeping hydrocarbons. This article provides guidance on best practices for seabed geochemical surveys based on both previously published laboratory experiments and the Marco Polo field calibration survey results. Previous SectionNext Section

FIELD OPERATIONS
Study Area
The Marco Polo field is approximately 175 mi (281 km) south of New Orleans, Louisiana, in blocks 563, 607, and 608 Green Canyon Gulf of Mexico in approximately 4000 ft (1219 m) of water (Figure 1). The Marco Polo field is located in a salt-bounded minibasin with petroleum production from suprasalt Miocene reservoir sands (Chaouche et al., 2004). Geochemical analysis of reservoir fluids by Anadarko Petroleum indicates that the medium-gravity oil originated from a marine type II organic matter type consistent with generation from the subsalt upper Jurassic source rocks (see Chaouche et al., 2004, for details). Areas of fluid movement from the reservoir to near surface had been identified on the western side of the field area (Green Canyon Block 607) by the presence of mud mounds (volcanoes), pockmarks (seabed craters), shallow acoustic blanking, and bright amplitudes (Dembicki and Samuels, 2007, 2008).

View larger version:

In this page In a new window

Figure 1 The Marco Polo field is approximately 175 mi (281 km) south of New Orleans, Louisiana, in blocks 563, 607, and 608 Green Canyon Gulf of Mexico in approximately 4000 ft (1219 m) of water.

Core Selection and Collection

Thirty-three cores were collected based on 23 targets and one location far from known seepage (Table 1). The core locations are displayed on a digital terrain map generated from highresolution multibeam bathymetry (Dembicki and Samuels, 2007) (Figure 2). Two types of features were targeted: within seep zone, sample location within major seepage zone based on high-resolution acoustic data; and near seep zone, sample location near feature with major seepage based on high-resolution acoustic data. A regional reference (local background reference) location was selected outside the minibasin for a baseline sample of recent sediment organic matter. More details on the core locations and targeted seismic features can be found in Dembicki and Samuels (2007, 2008). View this table:

In this window In a new window

Table 1 Core Description with Target Classification and Depth Information

View larger version:

In this page In a new window

Figure 2

The Marco Polo calibration cruise core locations on the surface terrain map, Gulf of Mexico, Green Canyon Block 607. Initial positioning of the vessel was done using shipboard global positioning system to match the location of features identified with autonomous underwater vehicle (AUV) data. The onboard Chirp subbottom profiler records were then compared with the AUV subbottom profiler records to confirm the targeted feature.The core samples were collected aboard the Fugro Seis Survey using a modified Kullenberg piston coring device (trough corer trap system) with a 6-m core barrel and a 6.7-cm internal diameter. The deep-water piston corer was launched and retrieved using a rotating A-frame high-speed hydroelectric dual winch. The 33 cores range in length from 0.5 to 5.3 m (1.617.4 ft) with an average recovery of 3.5 m (11.5 ft) (Table 1). Three core sections were collected from each sediment core: top (0.52.0 m [1.66.6 ft]), middle (2.0 3.0 m [6.69.8 ft]), and bottom (3.05.0 m [9.816.4 ft]). Each core section was subdivided in the shipboard laboratory and processed using the analytical protocols provided by the participating laboratories or established in the SGC laboratory studies: a 5-cm subsection placed in Gore Module sample container with a sorber; replicate 10-cm sample splits placed in a disrupter container (500-mL plastic container with a screw on sealing cap, built-in septum, and internal blades; see Abrams and Dahdah, 2010, for details) and a metal can for interstitial gas analysis; 3-cm sample placed in a plastic bag for microdesorption analysis; and replicate 10-cm sample splits were wrapped in aluminum foil for solvent extraction. Previous SectionNext Section

ANALYTICAL PROCEDURES
Sediment Gas Analysis
The conventional can headspace method is a nonmechanical procedure that uses high-speed shaking to release vapor-phase interstitial sediment gases into the can headspace (Bernard, 1978). A 500-mL metal can is filled with 170 mL of sediment, 160 mL of 3.5% NaCl in H2O solution, and the remaining headspace is flushed with N2 gas. The sample can with mud and processed water is stored frozen. The can is thawed for 24 hr before analysis, heated to 40C for 4 hr, and shaken vigorously using a conventional paint shaker. The headspace gas is collected by syringe and injected in the gas chromatograph for compositional analysis. See Bernard (1978) for more details. The disrupter sediment gas extraction method was designed as part of the SGC laboratory studies to capture interstitial gases not easily released by vigorous shaking (Abrams and Dahdah, 2010). The disrupter method uses a 165-mL sediment sample that is placed in a 500-mL disrupter chamber with 165 mL of saturated salt brine solution and remaining volume air headspace. The disrupter chamber has a fixed internal blade that breaks apart the sediment, releasing interstitial gases without crushing (Abrams and Dahdah, 2010). The disrupter with sample is frozen on the vessel, then shipped and stored frozen until analysis. The disrupter is thawed to room temperature 24 hr before analysis and shaken for 5 min using a high-speed unidirectional paint can shaker. A 0.2-mL disrupter headspace sample is collected by a syringe through the disrupter

cap septum at room temperature and injected into the GC inlet. See Abrams and Dahdah (2010) for more details. The bound gases are believed to be attached to organic and/or mineral surfaces, entrapped in structured water, or entrapped in authigenic carbonate inclusions and thus require a more rigorous procedure to remove (Horvitz, 1985; Bjoroy and Ferriday, 2002; Whiticar, 2002). The microdesorption bound sediment gas extraction method (Whiticar, 2002) uses a 300- to 400-mL bulk sediment sample that has been stored frozen in a plastic bag. A fixed weight of wet sediment sample (13 g) is placed in a reaction vessel, sealed, and evacuated. A small amount of saline water is added, and the sediment-water slurry is mixed using a vortex ultrasonic mixer. The interstitial gases are removed by vacuum. Phosphoric acid is added under reduced pressure where the sorbed gas is released to the vessel headspace. Potassium hydroxide is added before GC to reduce carbonate-generated carbon dioxide. The pressure is increased and sample aliquots of gas are collected for GC analysis. See Whiticar (2002) for details.

Gasoline-Range (C5C10+) Hydrocarbon Analysis

Two methods were used in the Marco Polo field calibration survey to evaluate sediment gasoline-range hydrocarbons, the Gore Module and disrupter HSPME. The Gore method evaluates a full range of hydrocarbons from C2 to C20+ using a specially engineered hydrophobic adsorbent encased in a microporous expanded Gore Module polytetrafluoroethylene membrane (Anderson, 2006). The Gore Module is placed in a special glass container with a designated volume of sediment and analyzed via thermal desorption coupled with mass spectrometry (MS). See the W. L. Gore and Associates Web page for additional details. (www.gore.com) The HSPME headspace sample is collected after the disrupter headspace gas analysis. A 1-cm fused silica fiber coated with 100-m-thick polydimethylsiloxane is inserted into the headspace for a 20-min extraction, then manually injected into the GC inlet for desorption. The SPME fiber is left in the GC inlet for 5 min. See Abrams et al. (2009) for details.

High-Molecular-Weight Hydrocarbon Analysis

A dried sediment sample is ground to a uniform size and an aliquot by weight is extracted using hexane in an automated extraction apparatus (Dionex ASE 200 Accelerated Solvent Extractor). Extracts are concentrated to a final volume of 8 mL using Zymark TurboVap II. The final extracts are submitted for hydrocarbon analysis by GC flame ionization detection (FID) and TSF. See Brooks et al. (1983) for additional information. Previous SectionNext Section

RESULTS AND DISCUSSION

Sediment Gas Analysis
Interstitial Gas Evaluation

The can headspace and disrupter data are summarized in Table 2 and plotted on a total hydrocarbon gas ( C1C5) versus wet gas fraction ( C2C5/ C1C5) plot (Figure 3). Both interstitial sediment gas extraction methods provide similar results. Three major groups exist for both interstitial sediment gas data sets using the classification scheme from Abrams (2005): background with total gas concentrations less than 10,000 ppm and low wet gas fraction (<0.05); fractionated with total gas concentrations less than 100 ppm and elevated wet gas fraction (>0.05); and anomalous with total gas concentrations greater than 10,000 ppm and wet gas fraction less than 0.1 (10%). The cutoffs are based on an examination of a worldwide surface geochemistry database (Abrams, 2005). View this table:

In this window In a new window

Table 2 Marco Polo Surface Geochemistry Calibration Can Headspace and Disrupter Interstitial Sediment Gas Data

View larger version:

In this page In a new window

Figure 3 The three major groups for Marco Polo can headspace and disrupter light hydrocarbon data using the classification scheme from Abrams (2005); background with total gas concentrations less than 10,000 ppm and low wet gas fraction (<0.05), fractionated with total gas concentrations less than 100 ppm and elevated wet gas fraction (>0.05), anomalous with total gas concentrations greater than 10,000 ppm and wet gas fraction less than 0.1 (<10%). Examination of the three sediment gas groups relative to the pre-survey core targets (within seep zone, near seep zone, and regional reference) reveals interesting observations. The regional reference core samples fall within the background and fractionated samples. The fractionated group was defined in Abrams (2005) to represent very low concentration samples (<100 ppm) with differential volatile loss (methane loss greater than the wet gases) that results in wet gas

enrichment. This is a common feature noted in most interstitial sediment gas seabed surveys (Abrams, 2005) and SGC laboratory experiments (Abrams et al., 2009). The within-seep-zone and near-seep-zone targeted cores fall within all three groups: background, fractionated, and anomalous. Note the relatively large number of cores collected from within seep zone and near seep zone based on AUV acoustic data that contain very low total interstitial hydrocarbon gas (<10,000 ppm). Several reasons exist why near-surface sediment cores collected within a seep feature contain low levels of total interstitial sediment hydrocarbon gas. First, the core samples were collected within the transition zone known as the zone of maximum disturbance (ZMD) (Abrams, 1992). Samples collected within the ZMD may be altered by microbial processes or pore water flushing (Abrams, 1996) unless the seepage volume and rate overwhelm the alteration processes (Abrams, 2005). A plot of core sampling depth versus total disrupter headspace interstitial hydrocarbon gas indicates that the ZMD is approximately 2.0 m (6.6 ft) based on the near-seep-zone samples with elevated total gases (>10,000 ppm) relative to sample depth (Figure 4). Note that the ZMD will vary for each petroleum seepage system (Abrams, 1996, 2005).

View larger version:

In this page In a new window

Figure 4 The plot of core sampling depth versus total disrupter headspace interstitial hydrocarbon gas indicates the zone of maximum disturbance (ZMD) is approximately 2.0 m (6.6 ft). The withinseep-zone samples have elevated total gas concentrations within and below the ZMD, whereas the near-seep-zone samples have elevated total gas concentrations only below the ZMD. The second and more fundamental issue is collecting cores within the chosen targeted feature. Sampling a relatively small target in 4000 ft (1219.2 m) of water with a gravity corer is very difficult. Many of the cores collected within the Marco Polo survey did not sample the targeted feature based on the geochemistry results. Abrams (1992) demonstrated the importance of sampling depth and hitting the target by collecting core samples from an anchored drill ship using shallow drilling technology on and off a known hydrocarbon seeping fault. The drill ship survey showed that the migrated thermogenic signal was quickly lost from deep sediment cores

collected 15 to 25 m (4992 ft) away from the migration pathway (leaky fault). The Marco Polo calibration data set reinforces the importance of good seismic data to define the seep feature and a core collection protocol with the ability to hit the targeted feature at 2+ m below the watersediment interface. The disrupter interstitial sediment gas composition (wet gas fraction and C1/C2 ratio) and Marco Polo field reservoir gases are different (Figure 5). Similar observations were made by Abrams and Dembicki (2006). The differences between reservoir and near-surface sediment gas compositions are most likely related to near-surface affects (phase fractionation and microbial alteration) as well as mixing with shallow microbial gases. Therefore, caution should be used when interpreting near-surface sediment gas compositions using interpretation charts designed for reservoir gases (Abrams, 2005).

View larger version:

In this page In a new window

Figure 5 The disrupter interstitial sediment gas composition (wet gas fraction and C1/C2 ratio methane/ethane) is different from the Marco Polo Field reservoir gases. Compound-specific carbon isotopic analysis (methane, ethane, and propane) of selected highconcentration macroseepage (anomalous) can headspace and disrupter extracted interstitial gases provide similar results except for core number 23C (Table 3). The disrupter 13C1 for sample 23C is 9 lighter, indicating in-situ microbial gas generation. The methane and ethane 13Cn values for the disrupter-extracted interstitial gases are similar to the 12,200-ft reservoir production gases, except propane (Figure 6). The disrupter and can headspace-extracted hydrocarbon gases have propane isotopic values heavier (13C3, 15.3 to 29.3) relative to the reservoir gases (13C3, 27.8 to 31.1) (Table 3; Figure 6). This is most likely related to nearsurface microbial alteration of propane. The preferential attack of propane was noted by James and Burns (1984) in reservoir gases, and microbial fractionation of near-surface gases is not uncommon for seabed sediment gases (Abrams, 1989). View this table:

In this window In a new window

Table 3 Marco Polo Surface Geochemistry Calibration Interstitial Sediment Gas (Can Headspace and Disrupter) Compound-Specific Carbon and Methane Hydrogen Isotopic Data

View larger version:

In this page In a new window

Figure 6 The methane and ethane 13Cn values for disrupter extracted interstitial gases are similar to the 12,200 ft (3719 m) reservoir production gases, except propane (C3) (does not include sample 23C). Microdesorption (Bound) Gases The microdesorption bound gas data (Table 4) are plotted on the same total hydrocarbon gas ( C1C4) versus wet gas fraction ( C2C4/ C1C4) evaluation plot using the same core designations (Figure 7). Note that the microdesorption sediment gases are reported in nanomoles per gram by weight, whereas the disrupter headspace gas is reported in parts per million by volume. View this table:

In this window In a new window

Table 4 Marco Polo Surface Geochemistry Calibration Microdesorption Sediment Gas Data

View larger version:

In this page In a new window

Figure 7 The bound gases display different trends from the interstitial gases. No bimodal distribution and total microdesorption hydrocarbon gas displays significant variability from the low- and highconcentration samples. Most have wet gas fractions less than 0.12 (<12%) with three notable exceptions: 0.48, 0.66, and 0.68. Most display higher gas wetness than the reservoir gases. The bound gases display a very different trend than the interstitial gases. No bimodal distribution is present and total microdesorption hydrocarbon gas displays significant variability from lowand high-concentration samples (Figure 7). Most of the samples have wet gas fractions less than 0.12 (<12%), with three notable exceptions: 0.48, 0.66, and 0.68 (4868%). Most of the microdesorption extracted sediment gases show higher gas wetness than the reservoir gases (Figure 7). The bound gases compound-specific isotopic ratios do not match the reservoir gases (Figure 8). The sediment-bound methane and ethane isotopes are as much as 7 heavier than the Marco Polo Field reservoir gases. This is not an uncommon observation for sediment-bound gases removed by acid extraction from both washed and closed-vessel adsorbed-hydrocarbon analysis (Abrams and Dahdah, 2010).

View larger version:

In this page In a new window

Figure 8 Most Marco Polo sediment-bound gases do not match the reservoir gas isotopic values. Bound gases are up to 7 heavier than the Marco Polo Field reservoir gases.

Gasoline-Range Hydrocarbon Analysis

The gasoline-range petroleum hydrocarbons comprise molecules with 5 to 12 carbon atoms (C5 C12) arranged in linear, branched, and cyclic aliphatic structures along with monoaromatic hydrocarbons, such as benzene, toluene, and o-, m-, and p-xylenes. This group of hydrocarbons is normally derived from thermogenic processes associated with a mature generating source rock unlike methane and ethane (C1 and C2) that could be derived from either thermogenic or microbial processes (Whiticar, 1999). The gasoline-range hydrocarbons are volatile and migrate within key oil migration avenues to the sediment surface and should be an important target for most surface geochemical surveys (Abrams et al., 2009). To date, few surface geochemical surveys attempt to evaluate the gasoline-range hydrocarbons in near-surface marine sediments. Conventional headspace light hydrocarbon analysis is not an effective method to evaluate the C6 to C12 hydrocarbons because of higher boiling points and low vapor pressures relative to the hydrocarbons gases (C1C5) (Abrams and Dahdah, 2010). Headspace Solid-Phase Microextraction The HSPME data are reported as the area sum of a single carbon number (SCN) within the main boiling point range detected using the SPME fiber. The unresolved complex mixture (UCM) is included in the area of each SCN (Abrams et al., 2009) (Table 5). A plot of the disrupter total hydrocarbon interstitial gas ( C1C5) versus HSPME SCN (Figure 9) displays moderate correlation between sediment interstitial gas and the concentration of gasoline-range hydrocarbons. View this table:

In this window In a new window

Table 5 Marco Polo Surface Geochemistry Calibration Headspace Solid-Phase Microextraction and Gore Module Gasoline Plus Data with Statistical Grouping Classifications

View larger version:

In this page In a new window

Figure 9 The headspace solid-phase microextraction (HSPME) data reported as sum of single carbon number (SCN) within the detection range of SPME fiber. The unresolved complex mixture is included in the area of each SCN. A plot of disrupter total hydrocarbon interstitial gas ( C1C5) versus HSPME SCN displays a moderate correlation between sediment interstitial gas and concentration of gasoline-range hydrocarbons with notable exceptions. The regional reference HSPME gas chromatograms contain very low overall signal responses (<10 total GC area) that are dominated by SPME fiber peaks (Figure 10A) (Abrams et al., 2009). In contrast, the high-concentration within-seep-zone HSPME gas chromatograms have an elevated overall signal response increasing with depth, significant resolvable compounds, depleted light end, and elevated baseline hump or UCM (Figure 10B).

View larger version:

In this page In a new window

Figure 10 (A) The headspace solid-phase microextraction (HSPME) data for regional reference core. (B) The HSPME data for the within-seep-zone core.

Compound distributions for the within-seep-zone chromatograms are very different to an unaltered reservoir oil whole-oil chromatogram. The HSPME within-seep-zone chromatograms contain very low n-alkanes (C5, C6, C7, and C8), aromatics (benzene, toluene, and xylenes), cycloalkanes (cyclopentane and cyclohexane), and cycloalkanes with one methyl group (methylcyclopentane and methyl-cyclohexane) but elevated isoalkanes and cycloalkanes with more than one methyl group. This unique compound distribution is commonly found in biodegraded oils (George et al., 2002), indicating that the macroseepage seabed gasoline-range hydrocarbons are rapidly and heavily altered (Abrams et al., 2009). Gore Module The Gore Module examines as much as 150 volatile compounds using cluster analysis and linear discriminate classification for defining petroleum and background character end members (Anderson, 2006). Three end-member groupings were defined by the Gore statistical evaluation: high aliphatic, medium aliphatic, and background (Table 5; Figure 11AC).

View larger version:

In this page In a new window

Figure 11 Three end-member groupings defined by the Gore Module data for the Marco Polo calibration data set: (AC) high aliphatic, medium aliphatic, and background. The Gore high aliphatic group contains the higher concentration (based on headspace gas and HSPME data) within-seep-zone samples, with two exceptions (Table 5). The key compounds that characterize this end member group include 2-methylbutane, 3-methylpentane, methylcyclohexane, cyclohexane, 1-octene, cyclopentane, c1314 dimethylcyclohexane, and ethylcyclohexane (Figure 11A). The Gore medium aliphatic group contains a mix of the within-seep-zone and near-seep-zone samples (Table 5). The key compounds that characterize this end member group include carbon disulfide, 2,4-dimethylpentane, 2,5-dimethylhexane, pristane, ethane, and cyclooctane (Figure 11B).

The Gore background group contains the regional reference samples as well as the lowconcentration within-seep-zone and near-seep-zone samples (Table 5). The key compounds that characterize this end member group include carbon disulfide, propene, ethane, octanal, decanal, propane, 1-butene, butane, and tetradecane (Figure 11C). A plot of Gore Module total reported hydrocarbons versus total disrupter HSPME SCN indicates that both gasoline-range sediment extraction methods provide similar results (Figure 12). Both methods provide strong gasoline-range signals within the interstitial sediment gas highconcentration within-seep-zone and near-seep-zone samples and minimal signal in the regional reference and interstitial sediment gas low-concentration within-seep-zone and near-seep-zone targeted cores. The Gore Module thermal desorption MS analysis provides greater compound specific information than could be achieved with the HSPME GC-FID approach. This could prove to be important to characterize seep origin (source facies, level of maturity, and/or seep to oil correlation). In addition, the desorption MS analysis can detect much lower concentrations than other seabed geochemical analysis methods. However, it could not discriminate the lowconcentration within-seep-zone and near-seep-zone samples any better than the disrupter headspace gas or HSPME gasoline-range analysis.

View larger version:

In this page In a new window

Figure 12 A plot of the Gore Module total reported hydrocarbons versus total disrupter headspace solidphase microextraction (HSPME) single carbon number (SCN) indicates that both gasoline-range sediment extraction methods provide similar results: strong gasoline-range signals within the interstitial sediment gas high-concentration within-seep-zone and near-seep-zone samples and minimal signal in the regional reference and interstitial sediment gas low-concentration withinseep-zone and near-seep-zone targeted cores. The 8, 11, 16, 19, and 25 within-seep-zone and near-seep-zone core shallow subsamples were classified by Gore in the background group, whereas the deeper subsamples fall in the medium aliphatic group. This observation reinforces the importance of sampling depth and placing the corer directly on the targeted feature.

High-Molecular-Weight Hydrocarbon Analysis

Extract GC: The sediment extract GC evaluation includes the chromatogram signature, total UCM, and total n-alkanes. Table 6 contains extract GC total UCM and total n-alkanes for the Marco Polo calibration samples. Samples with UCM values less than 25 g/g are considered to be background, whereas samples with UCM greater than 100 g/g are associated with migrated hydrocarbon seepage (Cole et al., 2001; Abrams, 2005). Six of the 15 cores collected at the within-seep-zone targets did not have elevated UCM (>100 g/g), indicating that the samples may not have hit the targeted feature (Table 6). One near-seep-zone sample contains elevated UCM (Table 6). View this table:

In this window In a new window

Table 6 Marco Polo Surface Geochemistry Calibration Extract Gas Chromatography and Total Scanning Fluorescence Data Sediments containing migrated thermogenic HMW hydrocarbons typically have GC-FID chromatograms characterized by a large unresolved complex mixture (UCM) with some discernible C15C32 n-alkanes and isoprenoids peaks, depending on the severity of microbial alteration (Figure 13A) (Brooks and Carey, 1986). All but one of the within-seep-zone samples have extremely high UCM values (>1000 g/g) and a sum of total alkanes less than 1. This is a common observation with sediment extract chromatograms and is related to the severe nearsurface bacterial alteration that has destroyed most of the resolvable normal alkane compounds very quickly. Sediments containing mainly the recent organic matter (ROM) signature with some migrated thermogenic signal will contain lower UCM and overprint of odd n-alkanes greater than C23 (Figure 13B) (Brooks and Carey, 1986).

View larger version:

In this page

In a new window

Figure 13 The Marco Polo Surface Geochemistry Calibration (SGC) extract gas chromatograms. (A) Upper: within seep zone. (B) Lower: regional reference. Extract TSF: The sediment extract TSF evaluation includes examination of the fluorogram signature and maximum fluorescence intensity (MFI). Samples with significant seepage require dilution before TSF analysis. The MFI values are adjusted by multiplying the measured MFI by the dilution factor to obtain a corrected MFI (Brooks et al., 1983). The extract TSF MFI ranges from 30,140 to 73,260,000 MFI units (Table 6) for the SGC phase III GOM calibration samples. Most of the samples collected in the within-seep-zone targets contain high-extract TSF MFI values (100,000 MFI units) (Cole et al., 2001; Abrams, 2005). Several near-seep-zone samples as well as one shallow regional-reference sample also have extract TSF MFI greater than 100,000 units. An extract GC total UCM versus TSF MFI plot demonstrates a relatively strong correlation between the two HMW screening tools (Figure 14), but examination of the TSF fluorogram signatures indicate a potential problem with the extract TSF method (Figure 15). The fluorogram signatures for a regional-reference sample with low-interstitial sediment gas, UCM, and TSF MFI; and within-seep-zone sample with highly elevated interstitial sediment gas, UCM, and TSF MFI should have very different shapes and locations of the maximum excitation wavelength (MFI Max Ex) and maximum emission wavelength (MFI Max EM), yet they are similar (Figure 15A, B). This indicates that the fluorogram shape, MFI Max Ex, and MFI Max EM may not assist in the identification of thermogenic seepage in marine sediments.

View larger version:

In this page In a new window

Figure 14 The extract gas chromatography (GC) total unresolved complex mixture (UCM) versus total scanning fluorescence (TSF) maximum fluorescence intensity (MFI) plot demonstrates relatively strong correlation between two high-molecular-weight screening tools for Marco Polo extract GC and TSF data.

View larger version:

In this page In a new window

Figure 15 Fluorogram signatures for regional reference sample with low interstitial sediment gas, unresolved complex mixture (UCM), and total scanning fluorescence (TSF) maximum fluorescence intensity (MFI); and within-seep-zone sample with highly elevated interstitial sediment gas, UCM, and TSF MFI should not have similar shapes and locations of maximum excitation wavelength (MFI Max Ex) and maximum emission wavelength (Max EM). (A) Regional reference target EGI-1 (322 cm [127 in.]): MFI = 56,480 and DIL 1:10. (B) Withinseep-zone target EGI-22 (60 cm [24 in.]): MFI = 67,845,000 and DIL 1:4000. Previous SectionNext Section

KEY OBSERVATIONS
Sediment Gases
The can headspace and disrupter extraction methods provide similar interstitial sediment gas data, indicating that the disrupter plastic container with screw cap and sealing gasket, built-in septum, and blades to break up the sediment did not provide significantly better results than the conventional can headspace method. Both methods provide highly variable gas compositions compared with the Marco Polo reservoir gases. Very few of the high-concentration within-seepzone samples have sediment gas compositions similar to the Marco Polo reservoir gases. In contrast, the can headspace and disrupter interstitial sediment gas compound-specific isotopes are similar to the Marco Polo reservoir gases, except for the propane carbon isotope value. The propane isotopic values are much heavier most likely because of preferential microbial alteration (James and Burns, 1984; Abrams, 1989, 2005). The bound gas extraction method did not provide gas compositions or compound-specific isotopes similar to the Marco Polo reservoir gases. This could be related to the bound gas removal process, which may fractionate the sediment gas sample (Abrams and Dahdah, 2010).

Proper identification of anomalous versus background interstitial (free) sediment gas is required to separate background samples that may seem to be thermogenic because of phase fractionation. Phase fractionation (preferential loss of methane relative to ethane plus gases) results in background low-concentration sediment gases having elevated wet gas fractions. The fractionated samples are commonly confused with migrated thermogenic hydrocarbons (Abrams, 2005) because of the higher relative wet gas component. Only the high-concentration samples are likely to be derived from migrated thermogenic gas. Compound-specific isotopic measurements are critical to confirm the thermogenic origin, assuming that isotopically distinctive changes are present. Many samples collected at within-seep-zone or near-seep-zone targets identified by conventional and high-resolution surface seismic data have only low concentrations (background) of interstitial gas. We believe that these samples did not hit the intended target or were collected within the ZMD.

Gasoline-Range Sediment Hydrocarbons

Both the HSPME and Gore Module extraction methods provide strong gasoline-range seepage signals for the high-concentration within-seep-zone and near-seep-zone samples and minimal to no signal in the regional reference and low-concentration within-seep-zone and near-seep-zone samples. The Gore Module thermal extraction combined with MS provides much greater compositional detail than the HSPME GC-FID method. The HSPME chromatograms show evidence for significant near-surface microbial alteration. The lack of an unaltered oil, even in zones of high flux macroseepage, leads us to believe that the rate of alteration is rapid. It is our belief that despite microbial alteration, gasoline-range hydrocarbons provide key information for a boiling point range not examined in most offshore surveys, and this type of data is very important to help identify subsurface hydrocarbon generation (Abrams et al., 2009).

High-Molecular-Weight Sediment Hydrocarbons

The within-seep-zone samples with elevated HMW hydrocarbons (macroseepage) have extremely high UCM values and a distinctive GC signature. The regional reference samples have much lower UCM with a different and distinctive GC signature. When concentrations of migrated HMW hydrocarbon are low relative to the in-situ ROM material, the identification of migrated thermogenic hydrocarbons is difficult. Reworked or transported hydrocarbons can be confused with locally migrated hydrocarbons (Abrams, 2005). Reworking and transported hydrocarbons have been identified within the Marco Polo Green Canyon area (Cole et al., 2001; Dembicki, 2010). The Marco Polo extract TSF data indicate that MFI measurements provide information on the presence of anomalous hydrocarbons, but TSF fluorogram shapes do not change with target type or other geochemical measurements (interstitial gas, gasoline-range hydrocarbons, or extract UCM). Edwards and Crawford (1999) demonstrated that a linear relationship between oil concentration and total fluorescence intensity can only be obtained between 0.10 and 10 ppm oil.

Within this range, fluorescence intensity is proportional to concentration, but outside this range, TSF fluorescence intensity (MFI) and shape can vary because of the dilution factor (hydrocarbon concentration). Other factors that affect TSF fluorescence intensity (MFI) and shape include absorbance (compounds in extract that can absorb either excitation or emitted light), quenching (energy can be transferred nonradiatively to coexisting molecules instead of being emitted as fluorescence), and secondary alteration (water washing, biodegradation, evaporative fractionation, and weathering). Previous SectionNext Section

CONCLUSIONS
The Marco Polo SGC and previous SGC laboratory studies provide a strong empirical data set to evaluate surface geochemical methods used by industry for marine surface geochemical surveys. The Marco Polo SGC calibration data set demonstrates the importance of targeted coring and sampling depth. To improve the detection of seabed migrated thermogenic hydrocarbon seepage, core samples should be collected along major migration pathways (cross-stratal leakage features) identified by conventional deep seismic and high-resolution sea floor imaging technology. Not all targeted cores will hit the designated feature, and thus, collecting replicates along key migration features is recommended. Collecting sediment samples below the ZMD is also important to reduce the transition zone alteration interference. Geochemical analysis should include a full range of hydrocarbon types; light hydrocarbon gases (C1C5), gasoline range (C5C10+), and HMW hydrocarbons (C15+). The two interstitial sediment gas extraction methods, can headspace and disrupter, provide similar results in both laboratory (Abrams and Dahdah, 2010) and field calibration studies. The interstitial sediment gas data should be plotted on a total hydrocarbon gas ( C1C5) versus wet gas fraction ( C2C5/ C1C5) chart to identify background, fractionated, and anomalous populations (Abrams, 2005). The Marco Polo survey anomalous interstitial sediment gases have variable gas compositions compared with the reservoir gases, thus, caution should be used when plotting seabed gases on conventional gas interpretation charts. The methane and ethane stable carbon isotopes from selected anomalous samples are similar to the Marco Polo reservoir gases. However, propane isotope values are much heavier, indicating that propane is more easily modified by in-situ microbial alteration. The microdesorption bound gases have gas compositions and compound-specific isotopes unlike the Marco Polo reservoir gases. The microdesorption sediment gases tend to have more wet gas fraction and highly variable methane carbon isotopes with heavier ethane carbon isotopes. The SGC laboratory studies indicate that prewashing to remove interstitial gases can have a major impact on bound gas results (Abrams and Dahdah, 2010). We do not recommend using bound gas extraction methods to evaluate subsurface hydrocarbons based on the Marco Polo and previous laboratory calibration studies reported in Abrams and Dahdah (2010).

The gasoline-range analysis provides a new range of hydrocarbons rarely examined in surface geochemical studies. Both the HSPME and Gore Module methods used in the Marco Polo calibration studies provide strong gasoline-range seep signals having useful information in the macroseepage (high-flux) and microseepage (low-flux) seep sites. The Gore Module thermal extraction combined with MS provides more compositional detail than the HSPME GC-FID method, which may be helpful to evaluate the seep hydrocarbon source and maturity. Extraction GC and TSF analyses provide information on the presence of HMW hydrocarbons in the Marco Polo calibration survey. The GC chromatogram signature and total UCM tracked the migrated thermogenic hydrocarbon macroseepage but did not work as well with the low-level microseepage samples. The TSF MFI data also directionally tracked migrated thermogenic hydrocarbon macroseepage and microseepage samples, but the fluorogram shape could not distinguish within-seep-zone and regional reference samples. We do not recommend extraction TSF to evaluate migrated HMW hydrocarbons in near-surface marine sediment seep surveys based on the above results. Follow-up studies by Dembicki (2010) with Anadarko examined extract saturate and aromatic fraction GC-MS to evaluate biomarker signature variability. Results of Dembicki (2010) demonstrate the importance of biomarker data to assist in detecting and interpreting lowconcentration petroleum seepage (microseepage). Biomarker data provide a means to characterize the ROM contribution and in turn allow for the identification of thermogenic hydrocarbons when the seep oil concentration is low relative to ROM. The above conclusions are based on the Marco Polo calibration study as well as previous laboratory calibration studies (Abrams et al., 2009; Logan et al., 2009; Abrams and Dahdah, 2010). They provide a framework to better understand how best to collect and extract migrated hydrocarbons from shallow-marine sediments and evaluate the results. However, it is also very important to integrate sediment hydrocarbon results with basin geology to fully understand how surface geochemistry observations relate to subsurface generation and potential entrapment (Abrams, 2005). A fully integrated evaluation provides the best petroleum systems interpretation model. Previous SectionNext Section

Acknowledgments
We thank the Surface Geochemistry Calibration (SGC) research project industry supporters and Energy and Geoscience Institute at the University of Utah. Ger van Graas (Statoil), Dennis Miller (Petrobras), Harry Dembicki (Anadarko), Neil Frewin (Shell), Andy Bishop (Shell), Brad Huizinga (ConocoPhillips), Angelo Riva (ENI), and Peter Eisenach (Wintershall) have all been extremely helpful during the various phases of the multiyear industry-funded research project. We thank Anadarko Petroleum for allowing the SGC research project to collect samples at the Marco Polo field and access to key seismic and geochemical information. We thank Harry Dembicki who arranged permission for the Marco Polo location site, chose and monitored the core locations, and participated in the cruise. We thank W. L. Gore and Associates and Taxon Biosciences, which provided staff and laboratory analysis at no cost. We thank Fugro and the

Seis Surveyor crew as well as the shipboard research staff Harry Dembicki (Anadarko), Matt Ashby (Taxon), Shuanglin Li (Qingdoa Institute of Marine Geology), Rowland Rincon (Peregrine Ventures), and Chris Reny (Peregrine Ventures). The figures were drafted by Jeffrey Massara with Apache Corporation. Reviews by Harry Dembicki, Barry Katz, and an unnamed reviewer were very helpful. The AAPG Editor thanks the following for their work on this paper: Harry Dembicki Jr., Barry J. Katz, and an anonymous reviewer.

Manuscript receivedAugust 13, 2010. Revised manuscript receivedNovember 22, 2010. Revised manuscript receivedJanuary 24, 2011. Final acceptanceMarch 21, 2011.

Michael A. Abrams is currently the manager of geochemistry for Apache Corporation. Before working with Apache, Michael was senior research scientist in the University of Utah's Energy and Geoscience Institute and senior research geochemist in Exxon Production Research Company. Michael's research interests include surface geochemistry, petroleum systems evaluation, and migration pathway analysis. Michael received his Ph.D. from the Imperial College London, a B.S. degree from George Washington University, and an M.S. degree from the University of Southern California. Nick Dahdah currently manages the organic geochemistry laboratory at EGI and has been a research scientist with the Institute since 1992, specializing in oil and source rock characterization. He previously worked as a well site geologist for the Natural Resources Authority in Jordan. He received a B.S. degree in geology from Southeast Missouri State University and an M.A. degree from the University of South Carolina. Previous Section

REFERENCES
1. 1. Abrams, M. A. , 1989, Interpretation of surface methane carbon isotopes extracted from surficial marine sediments for detection of subsurface hydrocarbons: Association Petroleum Geochemical Explorationist Bulletin, v. 5, p. 139166. 2. 1. Abrams, M. A. , 1992, Geophysical and geochemical evidence for subsurface hydrocarbon leakage in the Bering Sea, Alaska: Marine and Petroleum Geology Bulletin, v. 9, p. 208221, doi:10.1016/0264-8172(92)90092-S.

CrossRefWeb of Science 3. 1. Abrams, M. A. , 1996, Distribution of subsurface hydrocarbon seepage in near-surface marine sediments, inSchumacher, D., and Abrams, M. A., eds., Hydrocarbon migration and its near-surface effects: AAPG Memoir 66, p. 114. 4. 1. Abrams, M. A. , 2002, Surface geochemical calibration research study: An example of research partnership between academia and industry (abs.), in New insights into petroleum geoscience research through collaboration between industry and academia, abstract volume: Geological Society (London). 5. 1. Abrams, M. A. , 2005, Significance of hydrocarbon seepage relative to subsurface petroleum generation and entrapment: Marine and Petroleum Geology Bulletin, v. 22, no. 4, p. 457478, doi:10.1016/j.marpetgeo.2004.08.003. CrossRefWeb of Science 6. 1. Abrams, M. A., and 2. Dahdah, N. , 2010, Surface sediment gases as indicators of subsurface hydrocarbons: Examining the record in laboratory and field studies: Marine and Petroleum Geology, v. 27, p. 273 284, doi:10.1016/j.marpetgeo.2009.08.005. CrossRefGeoRef 7. 1. Abrams, M. A., and 2. Dembicki, H. , 2006, Variation in seabed geochemical signature relative to reservoir hydrocarbons (abs.): AAPG Search and Discovery article, http://www.searchanddiscovery.net/documents/2006/06088houston_abs/abstracts/abram s.htm (accessed July 7, 2011). 8.

1. Abrams, M. A., 2. Segall, M. P., and 3. Burtell, S. G. , 2001, Best practices for detecting, identifying, and characterizing near-surface migration of hydrocarbons within marine sediments: Proceedings of the Offshore Technology Conference, Houston, Texas, OTC Paper 13039, 14 p. 9. 1. Abrams, M. A., 2. Dahdah, N., and 3. Francu, E. , 2009, Development of methods to collect and analyze gasoline plus range (C5 to C12) hydrocarbons from seabed sediments as indicators of subsurface hydrocarbon generation and entrapment: Applied Geochemistry, v. 24, p. 19511970, doi:10.1016/j.apgeochem.2009.07.009. CrossRefGeoRefWeb of Science 10. 1. Anderson, H. S. , 2006, Amplified geochemical imaging: An enhanced view to optimize outcomes: European Association of Geoscientists and Engineers First Break, v. 24, p. 7781. 11. 1. Bernard, B. D. , 1978, Light hydrocarbons in marine sediments: Ph.D. thesis, Texas A&M University, College Station, Texas, 144 p. 12. 1. Bjoroy, M., and 2. Ferriday, I. , 2002, Surface geochemistry as an exploration tool: A comparison of results using different analytical techniques (abs.): AAPG Hedberg Conference, Vancouver, British Columbia, Canada, 2 p. 13. 1. Brooks, J. M., and 2. Carey, B. D. , 1986, Offshore surface geochemical exploration: Oil & Gas Journal, v. 84, p. 6672.

14. 1. 2. 3. 4. 5. Brooks, J. M., Kennicutt II, M. C., Bernard, L. A., Genoux, G. J., and Carey, B. D.

, 1983, Applications of total scanning fluorescence to exploration geochemistry: Proceedings of the Offshore Technology Conference, Houston, Texas, May 25, 1983, OTC Paper 4624, p. 393400 15. 1. Chaouche, A., 2. Gamwell, P., and 3. Brooks, J. , 2004, Surface geochemical evaluation of Green Canyon Block 608, Gulf of Mexico: A comparison of sea floor seeps and reservoir fluids (abs.): AAPG Annual Meeting, Dallas, Texas, April 1820, 2004. 16. 1. Cole, G., 2. Requejo, R., 3. DeVay, J., 4. Yu, A., 5. Peel, F., 6. Taylor, C. H., 7. Brooks, J., 8. Bernard, B., 9. Zumberge, J., and 10. Brown, S. , 2001, The deep-water GOM petroleum system: Insights from piston coring, defining seepage, anomalies, and background: 21st Annual GCS-SEPM Research Conference, p. 315342. 17. 1. Dembicki, H. , 2010, Recognizing and compensating for interference from recent organic matter and biodegradation during interpretation of biomarker data from sea floor hydrocarbon seeps: An example from the Marco Polo area seeps, Gulf of Mexico: Marine and Petroleum Geology, v. 27, p. 19361951, doi:10.1016/j.marpetgeo.2010.06.009. CrossRefGeoRef

18. 1. Dembicki Jr., H., and 2. Samuels, B. M. , 2007, Identification, characterization, and groundtruthing of deep-water thermogenic hydrocarbon macroseepage utilizing high-resolution AUV geophysical data, Offshore Technology Conference, Houston, Texas, April 30 to May 3, 2007. OTC Paper 18556, 11 p. 19. 1. Dembicki Jr., H., and 2. Samuels, B. M. , 2008, Improving the detection and analysis of seafloor macro-seeps: An example from the Marco Polo Field, Gulf of Mexico: International Petroleum Research Conference, Kuala Lumpur, Malaysia, December 35, 2008, IPTC Paper 12124. 20. 1. Edwards, D., and 2. Crawford, N. , 1999, UVF fluorimetry: Australian Geological Survey Organization Research Report. 21. 1. 2. 3. 4. George, S., Boreham, C. J., Minifie, S. A., and Teerman, S. C.

, 2002, The effect of minor to moderate biodegradation on C5 to C9 hydrocarbons in crude oils: Organic Geochemistry, v. 33, p. 12931317, doi:10.1016/S01466380(02)00117-1. CrossRefGeoRefWeb of Science 22. 1. Horvitz, L. , 1985, Geochemical exploration for petroleum: Science, v. 229, p. 812827. 23. 1. James, A. T., and 2. Burns, B. J. , 1984, Microbial alteration of subsurface natural gases accumulations: AAPG Bulletin, v. 68, p. 957960.

GeoRefAbstract 24. 1. 2. 3. 4. Logan, G. A., Abrams, M. A., Dahdah, N., and Grosjean, E.

, 2009, Examining laboratory methods for evaluating migrated high-molecular-weight hydrocarbons in marine sediments as indicators of subsurface hydrocarbon generation and entrapment: Organic Geochemistry, v. 40, p. 365375, doi:10.1016/j.orggeochem.2008.11.009. CrossRefGeoRef 25. 1. Whiticar, M. J. , 1999, Carbon and hydrogen isotope systematics of bacterial formation and oxidation of methane: Chemical Geology, v. 161, p. 291314, doi:10.1016/S0009-2541(99)00092-3. CrossRefGeoRefWeb of Science 26. 1. Whiticar, M. J. , 2002, Characterization and application of sorbed gas by microdesorption CF-IRMS: Near-surface hydrocarbon migration: Mechanisms and seepage rates: AAPG Hedberg Conference, Vancouver, British Columbia, Canada, April 710, 2002.