Hydrocarbons in Deep-Sea Sediments following the 2010 Deepwater Horizon Blowout in the Northeast Gulf of Mexico

The Deepwater Horizon (DWH) spill released 4.9 million barrels of oil into the Gulf of Mexico (GoM) over 87 days. Sediment and water sampling efforts were concentrated SW of the DWH and in coastal areas. Here we present geochemistry data from sediment cores collected in the aftermath of the DWH event from 1000 – 1500 m water depth in the DeSoto Canyon, NE of the DWH wellhead. Cores were analyzed at high-resolution (at 2 mm and 5 mm intervals) in order to evaluate the concentration, composition and input of hydrocarbons to the seafloor. Specifically, we analyzed total organic carbon (TOC), aliphatic, polycyclic aromatic hydrocarbon (PAHs), and biomarker (hopanes, steranes, diasteranes) compounds to elucidate possible sources and transport pathways for deposition of hydrocarbons. Results showed higher hydrocarbon concentrations during 2010-2011 compared to years prior to 2010. Hydrocarbon inputs in 2010-2011 were composed of a mixture of sources including terrestrial, planktonic, and weathered oil. Our results suggest that after the DWH event, both soluble and highly insoluble hydrocarbons were deposited at enhanced rates in the deep-sea. We proposed two distinct transport pathways of hydrocarbon deposition: 1) sinking of oil-particle aggregates (hydrocarbon-contaminated marine snow and/or suspended particulate material), and 2) advective transport and direct contact of the deep plume with the continental slope surface sediments between 1000-1200 m. Our findings underline the complexity of the depositional event observed in the aftermath of the DWH event in terms of multiple sources, variable concentrations, and spatial (depth-related) variability in the DeSoto Canyon, NE of the DWH wellhead.


Introduction
Sediments serve as repository systems for a large range of organic matter sources and hydrocarbons that can be used to assess historical impacts to the environment. In the northern Gulf of Mexico (GoM), the composition of organic matter deposited in deep-sea sediments is controlled by physical sorting of particles (regional hydrodynamics) and the off-shore movement of the less dense material from terrigenous sources transported by the Mississippi River [1,2]. In shallow coastal areas (e.g. estuaries) the terrigenous pool is primarily composed of fresh vascular plant detritus whereas in deeper areas off-shore it is comprised of highly altered organic matter from angiosperm grassland soils [3,4]. Compared to sediments in shallow areas, terrigenous derived matter in deep-sea sediments is typically recalcitrant [3]. Marine-derived organic matter content in sediments is greater in areas associated with high rates of primary productivity in the surface waters [4]. Other sources of sedimentary organic carbon are erosion of sedimentary rocks from the Mississsipi River drainage basin, and fossil fuels from natural oil seeps and petroleum exploration [3,5,6].
It is estimated that an average of 95,500 tons oil enters the GoM annually from natural seeps (73%), oil and gas extraction activities (3%), transportation activities (4%), and oil combustion byproducts (~16%) [7]. In comparison, the Deepwater Horizon (DWH) spill in 2010 released 4.9 million barrels of oil into the Gulf of Mexico (~699,700 metric tons) [8], an amount over seven times the average annual input of oil into the GoM [9]. The DWH blowout was unique not only for its size but also its depth at 1500 m below the sea surface. The released oil partitioned into oil drops, gas bubbles, and gas hydrates with accompanied partitioning of petroleum hydrocarbons into aqueous, gas and particulate phases. A mixture of soluble and insoluble petroleum hydrocarbons reached the sea surface, where it evaporated (5% of the leaked mass), formed sheens and slicks (10% of the leaked mass), was mechanically recovered (20% of the leaked mass) or burned (6% of the leaked mass) [10]. In addition, a lateral plume at~1000-1200 m depth formed (35% of the leaked mass) rich in water-soluble gases and compounds [10][11][12][13][14]. The final fate of the hydrocarbons from the subsurface plume is unknown, although dissolution and biodegradation have been proposed as important processes [11,14,15].
Following the blowout of the DWH drilling rig, an unusually large marine snow event was observed [16,17]. The large depositional event may have occurred by marine snow formation from surface oil slicks (containing~0.5 million barrels of liquid oil) and the subsurface deep plume (containing~1.7 million barrels of liquid oil). Despite the broad accounts of chemical composition of hydrocarbons at sea surface and depth in the GoM in 2010 [10] and their fate in the environment [18][19][20][21][22][23][24][25][26][27][28], a comprehensive discussion of possible transport pathways of oil into the deepwater sediment environment and a hydrocarbon inventory for 2010 is not available.
The primary objectives of this study are to contribute to the overall understanding of hydrocarbon geochemistry in deepwater sediments by providing information on the concentration and composition of sediments samples collected from the DeSoto Canyon, NE of the DWH; and to interpret these data within the context of the possible sources and transport pathways of hydrocarbons to the deep sea during the period of the study.

Sample collection
Sediment cores were collected in December 2010 and February 2011 on board Florida of Institute of Oceanogrpahy's (FIO) R/V WeatherBird II during oil spill response cruises WB1111 and WB1114. Three sediment-coring sites were located in the DeSoto Canyon in the northeastern GoM (Fig 1). Sites were located~56 km (DSH10 at 1520 m depth, 28.59°N, 87.53°W), 83 km (DSH08 at 1143 m depth, 29.07°N, 87.52°W) and~111 km (PCB06 at 1043 m depth, 29.06°N, 87.16°W) northeast of the DWH. These sites were chosen based on predictions of oil transport towards the opening of the canyon due to wind direction and the formation of the Loop Current during the summer of 2010 [29,30]. Due to the lack of studies conducted before 2010 in the studied sites (DSH10, DSH08, PCB06), a core was collected 240 km east of the DWH (NT1200 at 1200 m depth, 27°57.98' N, 86°1.38' W, 1200 m depth) to provide a control site representing an area that was outside of the impact area during the DWH event.
A multi-corer (Ocean Instruments MC-800) was used to collect sediment cores with minimal disturbance to the sediment-water interface and surficial sediment intervals. Whole cores were sliced on board or in the laboratory at the University of South Florida's College of Marine Science Paleo-Laboratory (USFCMS-PL) at 2 mm (0-20 mm) and 5 mm (>20 mm) intervals downcore. A modified version of the Engstrom [31] and Valsangkar [32] extrusion devices was used. The extruder consists of a basal metal plate holding a threaded steel rod with uniform threads calibrated (1 complete turn = 2 mm) to push the core at specific intervals vertically upward, where the sample was then sliced off the top. Extruded samples from cores designated for organic geochemistry analysis (bulk carbon and nitrogen content, stable isotopes, and hydrocarbon analyses) were stored in pre-combusted glass containers (450°C for 4hr) in a freezer (-20°C). Frozen samples were freeze-dried and ground using a mortar and pestle. Extruded samples from cores designated for analysis of short-lived radioisotopes geochronology ( 210 Pb, 234 Th) were weighed, freeze-dried, and weighed again to determine their wet mass and dry mass to calculate dry bulk density. A separate core was split and used for visual descriptions and X-ray photographs to confirm the absence of bioturbation that may compromise stratigraphy and geochronology. Sediments collected in December 2010 were analyzed down to a depth of 90 mm to include recently deposited and historical data, while sediments collected in February 2011 were analyzed down to 6 mm to analyze the most recently deposited sediment.

Geochronology and input calculations
Short-lived radionuclide geochronology samples were counted for 48 hours on a Canberra Series HPGe (High-Purity Germanium) Coaxial Planar Photon Detectors to determine excess 210 Pb and excess 234 Th activities for age dating. Samples were counted within 120 days of collection to account for the short half-life (24.1 days) of 234 Th. Raw activities were corrected for counting time, detector efficiency, as well as for the fraction of the total radioisotope measured, yielding activity in disintegrations per minute per gram (dpm/g) with error generally <5% of activity. Age dates were assigned to each sample analyzed using the Constant Rate of Supply (CRS) Model as described previously [33][34][35]. Organic carbon inputs to the sediment surface were calculated by combining compound concentrations with sediment density measurements and geochronology (age dates). Inputs were calculated for total organic carbon (TOC) integrated over specific 2 or 5 mm intervals corresponding to each year in the sediment cores analyzed.

Bulk carbon, nitrogen, and stable isotopes
Bulk analyses for total organic carbon (TOC), nitrogen (N), and carbon and nitrogen isotopic values (δ 13 C, δ 15 N) were carried out at USFCMS-PL. Prior to analysis of TOC and δ 13 C, preweighed subsamples were placed in glass containers and acidified (80% 1.0N HCl) to remove inorganic carbon [36]. Dried subsamples were then placed in silver capsules and analyzed using a CarloeErba 2500 Series 2 Elemental Analyzer coupled to a Thermo Finnigan Delta XL. All samples were analyzed in duplicate and data reported as the average (<1% difference between duplicates). The results are reported using conventional delta notation in permil (‰) units relative to the Vienna Pee Dee Belemnite (VPDB) standard. The precision for replicate analyses of external standards (NIST 8573, NIST 8574, NIIST 1570) was 0.3‰ for δ 13 C and 0.4‰ for δ 15 N (1σ, N = 100).
The aliphatic fraction was quantified in a gas chromatograph/flame ionization detector (GC/FID) by the external standard method in splitless injections of 5μL. A VF-1ms (15m x 0.25mm x 0.25μm) capillary column was used with a GC oven temperature programming of 80°C held for 0.5 min, then increased to 320°C at a rate of 10°C min -1 and held for 5.5 min. Injector temperature was set to 280°C. Identification and quantification of n-alkanes (nC 12 -nC 40 ) and isoprenoids pristine (Pr) and phytane (Phy) were conducted by comparing with reference standards. Branched alkanes and the unresolved complex mixture (UCM) were calculated using the mean response factors of n-alkanes. All samples were corrected for % recovery from spiked samples with surrogates and were generally 60-80%. Total aliphatic concentration was calculated as the sum of n-alkanes, isoprenoids, branched alkanes and UCM. Aliphatic compounds are expressed as sediment dry weight concentrations.
The aromatic fraction was quantified in a Gas Chromatograph/mass spectrometric detector (GC/MS) in full scan mode (m/z 50-550). Splitless injections of 1μL of the sample were conducted. A RXi 1 5sil column was used with a GC oven temperature programming of 60°C held for 8 min, then increased to 290°C at a rate of 6°C/min and held for 4 min, then increased to 340°C at a rate of 14°C/min, and held at the upper temperature for 5 min. The temperature of the MS detector was 250°C. Target  Concentrations of PAHs were calculated using response factors by comparison with a known standard mixture (16-unsubstituited EPA Priority Pollutants and selected congeners: Ultrascientific US-106N PAH mix, NIST 1491a) and were corrected for the recovery of the surrogate standard. When no commercial reference standard was available, compounds were quantitated using the response factor for an isomer. Therefore, the concentrations determined for many of the alkylated PAHs were semiquantitative. Recoveries from spiked samples included with each batch were generally within QA/QC criteria of 60-120%. Aromatic compounds are expressed as sediment dry weight concentrations.
Biomarkers (hopanes and steranes) were quantified using GC/MS/MS multiple reaction monitoring (MRM) on a Varian 320 triple quadrupole MS. Splitless injections of 1μL of the sample were conducted. Chromatographic separation of biomarker compounds were conducted using a RXi 1 5sil column (30 m x 0.25 mm x 0.25 μm) with a GC oven temperature programming of 80°C held for 1 min, then increased to 200°C at a rate of 40°C/min, to 250°C at 5°C/min, to 300°C at 2°C/min, to 320°C at 10°C/min, and held for 2 min. The GC was operated in constant-flow mode (1ml/min) with an inlet temperature of 275°C and a transfer line temperature of 320°C. Ion source temperature was 180°C and source electron energy was 70eV. Mass transitions targeted appropriate parent molecular ion masses on Q1 and monitored mass 191.2 (hopanes) or 217.0 (steranes) on Q3, with collision energy held at 10 volts throughout. Targeted hopanes (and relevant Q1 masses) included C 27 , C 28 , and C 29 norhopanes (370.5, 384.5, and 398.5, respectively), C 30 hopanes (412.5), and C 31 through C 35 homohopanes (426.5, 440.5, 454.5, 468.5, and 482.5), while sterane targets included C 27 -C 29 steranes and diasteranes (372.7, 386.7, and 400.7). An additional transition (376.7 to 221) was monitored to quantify the internal standard (cholestane 2,2,4,4 D4; CDN Isotopes). Argon at a pressure of 1 millitorr was used as a collision gas. Concentrations of biomarkers compounds were calculated using response factors by comparison with a known standard mixture (Hopane/Sterane calibration mix, Chiron, S-4436-10-IO) and the internal standard. When no commercial reference standard was available, compounds were quantitated using the response factor for the nearest available homologue in the same compound class. Concentrations were corrected for the recovery of the surrogate standard (d 50 -Tetracosane) for the F1 fraction. Recoveries from spiked samples included with each batch were generally within QA/QC criteria of 60-80%. Biomarker compounds are expressed as sediment dry weight concentrations.
Replicate hydrocarbon analyses were done in selected samples from the cores collected in 2010. Our depth resolution of the core samples allowed us to replicate hydrocarbon analyses at the 65-85 mm depth interval. The relative standard deviations (RSDs) of replicates (N = 4) for PAHs analysis were 10%, 17% and 19% for DSH10, DSH08 and PCB06 sites, respectively. RSDs of replicates (N = 4) for aliphatic analysis were 12%, 6% and 10% for DSH10, DSH08 and PCB06 sites, respectively. And, RSDs of replicates (N = 4) for biomarker analysis were 17%, 22% and 4% for DSH10, DSH08 and PCB06 sites, respectively. RSDs values for all compound groups were lower than the variability observed in concentration from surface sediment layers to downcore in all sites (see Results section).
Diagnostic ratios were calculated to discriminate hydrocarbon sources in the samples collected. Diagnostic ratios use isomer pairs that are abundant in different PAH sources but with similar dissolution and adsorption properties as they have comparable thermodynamic partitioning and kinetic mass transfer coefficients [39]. Typically, low molecular weight PAHs (LMW, containing 2-3 ring PAHs) are abundant in petrogenic sources while high molecular weight PAHs (HMW, containing 4-6 ring PAHs) are abundant in pyrogenic sources. However, some oils like the DWH oil contain HMW PAHs, but in moderate levels relative to LMW PAHs [14]. HMW PAHs can become more abundant due to loss of LMW PAHs during weathering processes (e.g. dissolution), therefore PAH diagnostic ratios must be interpreted with caution. We used the following diagnostic ratios: AN/(P+AN), pyrogenic index, PI: ∑(other 3-6 ring EPA priority pollutant PAHs) / ∑(5 alkylated PAHs), HMW/LMW, Parental/alkyl, (BAP+BGP)/HMW, and (ID+C)/HMW [40][41][42][43]. We also calculated diagnostic ratios using biomarkers due to their properties of source specificity and resistance to weathering and biodegradation [44][45][46]. We used the following diagnostic ratios: For comparison to potential hydrocarbon sources, diagnostic ratios were calculated for DWH crude oil (Macondo oil, MC252 block) obtained from British Petroleum (BP, sample No. SOB-20100622-084) and from crude oil standard reference material collected from the insertion tube that was receiving oil directly from the Macondo well (NIST 2779). We also used analyses from two sediment samples collected at the DWH wellhead that contain high concentration of 17α(H)21β(H)-hopane (60 ng g -1 ). In addition, we used GoM sediment for PAH ratios from sites with evidence of pyrogenic (HMW) PAH inputs reported by the Operational Science Advisory Team in 2010 [47].

Bulk measurements
In the DeSoto Canyon, recently deposited sediment sections were established in each sedimentary record indicating variable and larger deposition in 2010 compared to previous years ( Table 1, S1 Table). TOC input was from 2 to 4 times higher in 2010 compared to previous years (Fig 2), indicating that in 2010 TOC content in surface sediments is primarily the result of variation in water column-derived organic matter input.
The δ 13 C of organic carbon in the sediment samples showed a higher variation in 2010 at DSH08 (from -19‰ to -23‰), and in 2011 at DSH10 (-20‰ to -23‰) than in previous years (Fig 3, Table 1). δ 15 N values were similar down core in all sites (5.4 ± 0.1‰ at DSH10, 5.6 ± 0.1‰ at DSH08, and 4.2 ± 0.1‰ at PCB06) but higher in 2011 (up to~6.1‰ at all sites, Fig 3). C:N is higher in 2010 at DSH08 and PCB06, and in 2011 at DSH10 compared to previous years, although downcore C:N ratios showed large variations for all sites (Fig 3). The Mean general trends observed in the bulk measurements indicate higher variation of organic carbon sources and inputs in 2010 at DSH08 and PCB06, and 2011 at DSH10 compared to previous years.

Aliphatics in sediments
Total concentration of aliphatics in the DeSoto Canyon varied from 5 to 337 μg g -1 . Higher concentrations were observed in 2010 for shallower sites (up to 337 and 35 μg g -1 for DSH08 and PCB06, respectively; Fig 3) and in 2011 for the deeper site (up to 46 μg g -1 for DSH10; Fig  3). There was a sharp increase in aliphatic concentration in most sediment intervals corresponding to 2010 (in DSH08 and PCB06) and 2011 (in DSH10 and DSH08). The intermittent increase in concentration observed in the surface sediments is indicative of a recent sedimentary depositional event (Fig 1). This is supported by the trend observed in UCM (unresolved complex mixtures of hydrocarbons) ( Table 1). Elevated concentrations of UCM are generally attributed to petroleum hydrocarbons in contaminated sediments [48][49][50]. The presence of UCM was observed in all sites in nearly all years, but higher concentrations of at least one Mean order of magnitude were found in some sediment intervals during 2010 (Table 1). Moreover, there were still high concentrations of UCM in 2011 for two sites (DSH10 and PCB06). Long-chain (C 26) and short-chain (C 25) n-alkanes varied in concentration (Table 1). In a consistent pattern, long-chain n-alkanes accounted for >80% of the total n-alkanes in all years for all sites indicating dominance of terrestrial inputs. Similarly, calculations of the value of CPI 14-35 indicated a predominance of odd-numbered carbon chains in all years (Table 1) typical of compounds from terrestrial vascular plants [50][51][52].
Isoprenoids (Pr and Phy) are more resistant to degradation than n-alkanes resulting in a decrease of the ratios C 17 /Pr and C 18 /Phy in weathered samples [52][53][54]. Lower C 17 /Pr and C 18 / Phy ratios were observed in more recently deposited sediments (corresponding to 2010) than in downcore sediments, probably due to a rapid deposition of weathered hydrocarbons  (Table 1). In 2011, isoprenoid ratios changed for all sites suggesting a different source of organic matter. In some 2011 samples high ratios are present (>2) indicating a more biogenic input to the surface sediments. In downcore sediment layers corresponding to 2006, high values were observed in C 18 /Phy and low values in C 17 /Pr for the shallower sites (DSH08 and PCB06) suggesting a high deposition of biogenic material post Hurricane Katrina.
The general trends observed in the aliphatic fraction indicate a mixture of organic carbon sources and increased concentrations in 2010 (DSH08 and PCB06) and 2011 (DSH10) in contrast to previous years. In contrast, similar organic carbon sources and low concentrations were observed in all depth intervals (2011 to 1990) in the control site, NT1200 (Table 1, S1  Fig). Aliphatic ratios in the control site are similar to data from prior years to 2010 observed in the other study sites (Table 1).

PAHs in sediments
Total PAH concentration in the DeSoto Canyon varied from 70 to 524 ng g -1 . Higher concentrations were observed in 2010 for the shallower sites (up to 524 and 329 ng g -1 for DSH08 and PCB06, respectively) and in 2011 only for the deeper site (up to 373 ng g -1 for DSH10) (Table 2, Fig 2). Also, a disparity between TOC input and PAH concentrations of LMW and  Table 2. HMW compounds was observed in 2010 and 2011 (Fig 2). For PCB06, LMW PAHs were higher at the 4-8 mm interval followed by an increase of TOC flux at the 0-4 mm interval. For DSH08, LMW PAHs showed two peaks (the 0-2 mm and 4-6 mm intervals) concurrent with a high TOC flux (0-8 mm). In contrast, LMW PAHs at DSH10 show an increase in 2011 at the 2-4 mm interval when TOC flux is low. DSH10 was the only location with higher HMW than LMW PAH concentrations and the HMW PAHs consistently increased from the end of 2010 to 2011 when TOC flux was high (Fig 2). In contrast, for the control site (NT1200) LMW and HMW PAHs were consistently low in all depth intervals ( Table 2, S1 Fig).
The intermittent increase in concentration observed only in the surface of all sediment cores except NT1200, indicates a recent sedimentary depositional event in the DeSoto Canyon. This is supported by the notion that PAHs are known to be relatively persistent in sediments due to their hydrophobicity and particle adsorption affinity [55]. PAH preservation in the sedimentary environment occurs on a decadal time scale [56,57] while LMW PAHs have a half-life of days (e.g. anthracene) [56]. In our study, we observed some LMW PAH compounds with higher levels downcore in the years between 2004 to 2009 (Table 2) suggesting that degradation of these compounds may be low at these sites.
Additionally, the relative composition of PAHs by their number of rings showed a larger variation in 2010 and 2011 than in previous years. DSH10 showed a distinct distribution of PAHs in 2011 (~15% increase in 4-rings, up to 25% increase in 3-rings,~10% decrease in 6-rings; Fig 2) with diagnostic ratios suggesting a mixed source of pyrogenic (abundant HMW PAHs) and petrogenic PAHs (Table 3). Also, there were differences in the relative composition of PAHs at DSH08 in 2010 and 2011 (~16% increase in 3-rings,~10% decrease in 4-rings; Fig  2), but PAHs present were primarily from a petrogenic origin (Table 3). There was also a distinct variation in the composition of PAHs at PCB06 in 2010 (up to~20% increase in 2-rings, 10% increase in 3-rings, and~15% decrease in 4-rings; Fig 2) also primarily composed of PAHs from a petrogenic origin (Table 3).

Hopanes in sediments
Total concentration of biomarkers in the DeSoto Canyon varied from 41 to 776 ng g -1 (Fig 4). Higher concentrations of hopanes (up to 317 and 457 ng g -1 for DSH08 and PCB06, respectively), steranes (up to 247 and 274 ng g -1 for DSH08 and PCB06, respectively), and diasteranes compounds (up to 105 and 115 ng g -1 for DSH08 and PCB06, respectively) were observed in 2010-2011 for the shallower sites. Higher concentrations of biomarkers at the deeper site (DSH10) were observed only in 2011 (up to 139, 116 and 44 ng g -1 for hopanes, steranes and diasteranes, respectively). In contrast, for the control site (NT1200) biomarkers were consistently low in all depth intervals (up to 60, 20.6 and 5.2 ng g -1 for hopanes, steranes and diasteranes, respectively; S1 Fig). Higher concentrations were only observed in the surface intervals of all sediment cores except NT1200, indicating a recent sedimentary depositional event in the DeSoto Canyon, as suggested as well by aliphatic and PAH compounds (Fig 1). A less pronounced peak was observed downcore (at~2006) in the shallower sites in the DeSoto Canyon indicating a depositional event post Hurricane Katrina. Elevated concentrations of biomarkers are generally attributed to petroleum hydrocarbons in contaminated sediments due to their recalcitrant nature and source-specific compound distribution. The distribution profiles of biomarkers in our samples and in GoM oils are dominated by 17α(H), 21β(H)-hopane compound. In addition, C27-C29 diasteranes compounds are abundant with the 20R isomers more abundant than the 20S isomers. Typical biomarker ratios used in forensic analysis have shown to be useful in identifying the DWH event as the sourceoil in samples collected from beaches (sand, rocks, tar balls) [45,46]. However, some biomarker Table 3. PAH diagnostic ratios for the cores collected in 2010 (interval: 2010 to~2002) and 2011 (only 2011 data) at the studied sites, DWH oil, and a sediment sample (GoM sed.) reported by OSAT 1 (2010 Averages shown as arithmetic mean ± CI. (1) GoM sediment data from sites with evidence of pyrogenic input collected in 2010 (OSAT 1, 2010). doi:10.1371/journal.pone.0128371.t003 Hydrocarbons in Deep-Sea Sediments, Northeast Gulf of Mexico compounds can degrade compromising their use to some extent for source oil identification [45,52]. We found diagnostic ratios of known recalcitrant biomarkers (Table 4) to vary between crude oil (MC252 oil, NIST standard) and contaminated sediment samples (see the method section for details). Vertical transport of spilled oil through the 1.5 km water column and horizontal transport to the study sites weathered biomarker compounds, as observed in a deepwater coral community [28]. Thus, biomarker ratios from deep-sea samples should be interpreted with caution. Most ratios determined for 2010-2011 samples lie within the values of MC252 oil and contaminated sediment samples (with ±20% analytical uncertainty), indicating a match with DWH oil (Table 4). Specifically, DSH10 showed a match with MC252 oil, DSH08 matched both with MC252 oil and contaminated sediments, and PCB06 mostly matched MC252 oil. In contrast, most ratios determined for pre-2010 samples did not match MC252 oil or the contaminated sediment samples, indicative of other sources for biomarkers. Similarly, the control site (NT1200) did not match MC252 oil or the contaminated sediment samples in all depth intervals (Table 4).

Hydrocarbon concentrations
The results from the sites in the DeSoto Canyon indicate that higher amounts of hydrocarbons reached deepwater sediments in the study area in 2010-2011 for the shallower sites (DSH08 and PCB06) and in 2011 for the deeper site (DSH10) than in previous years. Specifically, aliphatic, PAH and biomarkers concentrations were higher in 2010-2011 (up to 337 μg g -1 , 525 ng g -1 , and 776 ng g -1 , respectively) compared to previous years (up to 50 μg g -1 , 320 ng g -1 , and 523 ng g -1 respectively) and the control site (up to 30.4 μg g -1 , 45.0 ng g -1 , and 85.9 ng g -1 respectively) ( ) [58] relative to the elevated concentrations we observed in 2010 and 2011. Also, elevated concentrations in 2010-2011 are close to historical data from shallow water sites adjacent to and in the Mississippi River Delta (PAHs concentration decade 2000:~600 ng g -1 , decade 1980:~800 ng g -1 ) [59,60] but lower than urbanized and industrialized areas worldwide (only parental PAH concentration:~1200-11000 ng g -1 ) [61,62].

Hydrocarbon sources
Organic matter in the continental shelf sediments of the GoM is derived from marine primary productivity, terrestrial and aged soil from watershed runoff from the Atchafalaya and Mississippi Rivers, and coastal plants debris and resuspended sediments [4,67,68]. Most of the suspended particulate material is deposited within 30 km off the Mississippi River and a gradient of terrestrial vs. marine sources is observed towards deeper environments [68]. At~200 km from the Mississippi River delta in the DeSoto Canyon terrestrial inputs continue to dominate [69] as observed in the sediment cores from our study sites with n-alkanes typical of terrestrial sources (Table 1). Transport and deposition of pyrogenic PAHs also occur along with sediment deposition near the Mississippi River mouth and transported elsewhere in the GoM [68,70], as observed downcore in our samples ) with HMW PAH concentrations higher or similar to LMW PAH concentrations. Petroleum derived hydrocarbons in GoM sediments are consistently present and are suggested to originate from natural seeps, gas hydrate deposits, and oil exploration [6,[71][72][73][74]. Natural seeps can account for most of the annual input of oil in the GoM (~73%) [7], and their significant presence in the environment can be detected when depleted stable carbon isotope values are observed in sediments (-24.1‰) [75]. Downcore samples prior to 2010 in our study ) indicate sedimentary δ 13 C values of -20.6 ± 0.1‰, typical of a mixed organic carbon source including terrestrial-derived organic carbon from the Mississippi river drainage basin found in deep water sediments [67,76]. Previous studies indicate as well absence of seeps at specific sites in the DeSoto Canyon [58]. We conclude that during the period covered in our study  concentration and composition of hydrocarbon compounds at our study sites were not influenced directly by hydrocarbon inputs from natural seeps.
Petroleum derived hydrocarbons, distinct from DWH oil, were observed downcore (pre-2010) in all sites indicating the constant presence of oil at lower levels that enters the GoM annually through natural and anthropogenic activities [7]. Recalcitrant compounds such as biomarkers are useful for estimating background conditions in the GoM sedimentary environment. Mean background concentrations were 13 ± 3 ng g -1 for 17α(H)21β(H)-hopane, and 96 ± 18 ng g -1 for total biomarkers (N = 19) at~1500 m depth, and 40 ± 7 ng g -1 for 17α(H)21β (H)-hopane, and 301 ± 46 ng g -1 for total biomarkers (N = 37) at 1000-1200 m depth in the DeSoto Canyon. Depth variability rather than location in the DeSoto Canyon has a larger influence on the magnitude and time of hydrocarbon deposition during the period covered in our study. Similarly, depth-related gradients of deposited organic matter have been observed elsewhere in the northern GoM [1,3].
Elevated concentrations of aliphatic, PAH and biomarker compounds were observed in most sediment depth intervals corresponding to the period 2010-2011. These compound groups indicate a mixture of hydrocarbons sources deposited during 2010-2011 that includes terrestrial, planktonic and weathered oil. Long-chain n-alkanes distribution suggests predominance of terrestrial sources, while higher concentrations of UCM, PAHs, and biomarkers indicate oil-derived hydrocarbons (Figs 2, 3 and 4). Furthermore, diagnostic ratios of biomarker compounds suggest DWH oil reached the seafloor in the DeSoto Canyon during 2010-2011 (Table 4). However, highly weathered deposited hydrocarbons as indicated by the isoprenoids ratios (Table 1), complicate the use of diagnostic ratios for source oil identification. Spilled oil from the DWH wellhead underwent weathering during both lateral and vertical transport primarily from partitioning of the more soluble compounds into the water column [10][11][12]14] and enhanced microbial activity [16,17,77]. Nevertheless, the time series data revealed elevated concentrations of hydrocarbons (from more degradable to recalcitrant compounds) at specific intervals in the surface of the seafloor, indicating a depositional event at all sites in the DeSoto Canyon during 2010-2011, post DWH event.

Transport pathways of hydrocarbons
Two mechanisms of hydrocarbon deposition to the seafloor have been proposed: a sedimentary depositional pulse driven by the formation and rapid settling of contaminated particles ("flocculent blizzard" hypothesis), and the direct contact of the deep plume with the continental slope surface sediments at depths between 1000-1200 m ("bathtub ring" hypothesis). In our study, we found several different lines of geochemical evidence that support each of these hypotheses and explain how hydrocarbons reached the sedimentary environment in the DeSoto Canyon during 2010-2011.
Aggregation of suspended particulate material with dispersed crude oil and dissolved hydrocarbons in the water column promotes transport of spilled oil in the environment and toxicity in sediments. This phenomenon of oil-mineral interaction and sedimentation has been studied and documented since the 1970's in the Exxon Valdez and Arabian Gulf oil spills and in laboratory experiments [78,79]. It is believed that oil-mineral aggregates are insignificant in locations far from coastal areas due to minimal suspended particulate material concentrations [78]. However, in 2010 the increased release of water from the Mississippi River with nutrients and fine-grain clays coupled with the long trajectory of DWH oil to the sea surface, and the abundance of oil in the water column (~5.5x10 6 kg of hydrocarbons at the deep plume) [10] may have enhanced oil-mineral aggregation and enhanced the observed depositional event. Additionally, the increase in microbial activity after the DWH blowout [20,77] stimulated the formation of a large marine snow event in the GoM that also may have promoted the formation of oil-mineral aggregates [16,17]. This explains the presence of flocculent material on top of deep corals [28], impact on benthic communities [27], and hydrocarbon inputs to the seafloor (this study).
The observed marine snow during the summer of 2010 in the GoM consisted of phytoplankton cells, bacteria, and oil-particle aggregates [16]. Our sediment analyses from the Desoto Canyon indicate deposition of oil-particle aggregates (-23.1‰ at DSH10 in 2011, and -22.5‰ at DSH08 in 2010), biogenic material (abundant n-alkanes C25 in all sites, dominance of marine inputs at DSH08 in 2010 with δ 13 C-TOC = -19.1‰), and degraded oil and organic matter (all sites) at specific depth intervals corresponding to 2010-2011. Chronological comparison (before and after the DWH event) of source inputs using aliphatics composition indicates little variation, therefore high inputs of background algal and terrestrial biomass in 2010-2011. As hypothesized, the sinking of oil-particle aggregates ("flocculent blizzard") may have purged the water column from particles and DWH oil, leading to the deposition of natural and petroleum derived hydrocarbons to the seafloor. Additionally, oil-particle aggregates in 2010-2011 may have formed with hydrocarbons different from crude or dissolved oil in the water column. We observed elevated concentrations of HMW PAHs at DSH10, the deepest site in 2011 (Table 2). Although HMW PAHs (pyrogenic) that are associated with atmospheric deposition and riverine inputs have been observed in sediment cores from other regions in the GoM [59,70] PAH source identification in deepwater sediments in the GoM are yet to be fully understood. PAH composition at DSH10 site does not indicate atmospheric deposition during 2010-2011 based on the ratio (BaP+BGP)/HMW (Table 3) used as an indicator for traffic-related pollution [80,81]. Therefore, there are three possible sources that can explain the elevated concentrations of HMW PAHs in 2011: 1) the large extent of the surface oil slick in 2010 was exposed to relatively high summer temperatures (25-30°C) [82] that enhanced the evaporation of the oil released [10]. Moderate evaporation of oil mousse was observed with preferential loss of LMW PAHs and <C14 n-alkanes [82] leaving a heavier residue (high in HMW PAHs) in the environment that formed oil-particle aggregates that sink. 2) Another possibility is deposition of burned products produced by in situ combustion of surface oil slicks during the summer of 2010 (6% of the leaked mass or 222,000-313,000 barrels) [10,83]. Burn residues are known to leave residues rich in HMW PAHs [40]. To our knowledge, there are no specific PAH ratios that identify burned residues from in situ combustion of DWH oil. 3) Diesel exhaust from thousands of vessels operating in the area during summer of 2010 (up to~6000 vessels in July 2010) may have been a source of HMW PAHs to the aquatic environment based on the diagnostic ratio (ID+C)/HMW [80,81] observed at DSH10 site (Table 3).
Poor correlation between TOC fluxes and hydrocarbon concentrations, and differential PAH abundance between HMW and LMW compounds at PCB06 site located at 1043 m depth (Fig 2) suggest the sinking of oil-particle aggregates was not the only transport pathway in 2010. Another possible depositional mechanism is the advective transport and subsequent contact of the deep plume with the continental slope surface sediments ("bathtub ring" hypothesis). At PCB06 site, specific sediment intervals in 2010 showed a dominance of LMW over HMW PAHs (Fig 2). Typically in surface oil spills, LMW compounds are lost through rapid evaporation limiting its dissolution into the water column, contrary to observations in the DWH deep plume [10,14]. Insoluble non-volatile compounds formed surface slicks while more soluble compounds were entrained in the subsurface plume [10]. Therefore, the concentration of LMW PAHs is a promising utility to track plume-derived hydrocarbons to deep-sea sediments, contrary to only using one insoluble compound not abundant in the plume such as 17α(H)21β(H)-hopane to describe this transport pathway [84]. Because there was no observed increase in sediment deposition rates at the time LMW PAHs were higher in concentration in PCB06 (4-8 mm interval in 2010), a direct contact of the deep plume with the surface sediment might have occurred. This is supported by an oil droplet model that suggests the deep plume impinged on the continental slope in the DeSoto Canyon [85] near the two shallower sites in our study (PCB06, DSH08). In addition, it is possible that the two proposed mechanisms of hydrocarbon deposition can occur simultaneously. For example, DSH08 at 0-2 mm in 2010 presented increased concentration of LMW PAHs as well as TOC rates indicating a possible hydrocarbon input from sinking oil-particle aggregates and from the direct contact of the deep plume with the surface sediments (Fig 2).

Conclusion
The analyses of DeSoto Canyon sediments show the contribution of the DWH event to sediment hydrocarbon content and composition, and provides insight into the transport pathways for hydrocarbons to the seafloor and their fate during 2010-2011. Higher hydrocarbon concentrations, and associated mixed hydrocarbon compositions from various sources (including biogenic n-alkanes, and petrogenic and pyrogenic PAHs) were observed during 2010-2011 in the DeSoto Canyon. Evidences were found for two hypothesized transport pathways of hydrocarbons: sinking of oil-particle aggregates and the direct contact of the deep plume with the sediment surface. Our results underline the complexity of hydrocarbons deposition due to weathering during transport, multiple sources, variable concentrations, and spatial (depth-related) variability in the DeSoto Canyon in 2010-2011. Further understanding of the interplay between the water column and sedimentary environments and the relative proportion of compounds deposited is critical for predicting the long-term fate of hydrocarbons from deep-water blowout events.