Intertidal sediments (mainly quartz sand) were collected from the southeastern Mediterranean coast (Lat. 32° 47′52N; Lon. 34° 57′20E) in the winter of 2019. The sediments were transferred to nine benthocosm tanks (∼140 L each), and the thickness of the sediment layer in each benthocosm tank was ∼20 cm. Ambient seawater was added to the benthocosm tanks so that sediments were submerged (water height was ∼25 cm above the sediments; Figure 1). To maintain ambient sea surface temperatures, the tanks were immersed in an elongated seawater tank (5 m³) at the Israel Oceanographic & Limnological Research (IOLR) Institute (southeastern Mediterranean coast). Porewater was collected using a custom-made extraction apparatus placed in the middle of each tank by a tube connected to a syringe. Following 3 days of acclimation, the nine benthocosms were assigned to three treatments (scattered randomly in the raceway pool): 1) crude-oil, 2) gas-condensate and 3) unamended control. A ∼500 µm thick layer of cru oil or gas condensate was established at the water surface, representative of a moderate-high pollution scenario based on the Israeli Ministry of Environmental Protection criteria (www.sviva.gov.il). The top sediments and porewater were sampled from each experimental benthocosm once every 7–14 days for >6 months (n = 17 samplings).

Pore-water samples were extracted using an acid-clean syringe and tubing through the collection apparatus (Figure 1), filtered through an 0.45 µm syringe filter, and added to acid-washed 20 ml plastic scintillation vials. The samples were kept at −20°C until analysis. NO₂ ⁻ + NO₃ ⁻ (nitrate + nitrite) and orthophosphate (PO₄ ³⁺) were determined using a segmented flow Seal Analytical AA-3 system. All measurements were well above the limit of detection; 80 μmol kg⁻¹ for NO₂ ⁻ + NO₃ ⁻ and 8 μmol kg⁻¹ for PO₄ ³⁺ (Sisma-Ventura et al., 2021).

TOC was measured by wet oxidation as in Gaudette et al. (1974) with minor modifications following Astrahan et al. (2017). Briefly, samples were oxidized with potassium dichromate and sulfuric acid after which the excess dichromate was potentiometrically titrated with ammonium ferrous sulfate using a Metrohm 785 DMP titrino auto-burette. The level of quantitation (LOQ) is 0.06%.

Total PAHs were analyzed using a GC-MS according to EPA8270 method. The LOQ for each different PAH measured was 5 ng gr⁻¹, while the detection levels varied between 0.46 and 1.4 ng gr⁻¹, depending on the molecule in question.

Samples were collected from the upper 1 cm of the benthocosm sediment (∼3 gr), placed in sterile plastic tubes containing 6 ml of pre-filtered seawater (0.2 µm) and were immediately fixed with a glutaraldehyde solution (1% final concentration, Sigma, G7651). At the lab, the chelating agent sodium pyrophosphate (final concentration 0.01 M) and the detergent Tween 20 (final concentration 0.5%) were added to reduce adhesion of bacterial cells to the sediment grains. The tube was then mixed for 30 min using vortex at 700 rpm before being placed in an ultrasonic water bath (Symphony, VWR) for 1 min to disperse the attached bacterial cells into the liquid phase. The tubes were left overnight at 4°C and subsamples were either stained with 1 μl SYBR green (Applied Biosystems cat #S32717) per 100 µl of sample in the dark (10 min) for heterotrophic bacterial counts or left unstained for picophytoplankton quantification. The samples were analyzed by an Attune® acoustic focusing flow cytometer. Cell abundance was normalized to the sediment’s dry weight.

Samples were collected from the upper 1 cm of the benthocosm sediment (∼10 gr), transferred to glass scintillation vials (20 ml) and stored at −20°C until analysis. Chlorophyll. a was extracted from the sediments using 90% acetone overnight at 4°C. The extracts were filtered through a 0.45 μm Teflon syringe filter and transferred into new glass vials. Chlorophyll. a concentrations were determined from the extracts by the non-acidification method (Welschmeyer, 1994) using a Turner Designs Trilogy® fluorometer at 436 nm excitation filter and 680 nm emission filter. Liquids were decanted from the sediment samples to determine their weight, and the values were used to calculate the amount of chlorophyll.a per gram of sediment.

Samples were collected from the upper 1 cm of the benthocosm sediment (∼3 gr), re-suspended in 3.5 ml of pre-filtered seawater (0.2 µm) and spiked with 500 nmol L⁻¹ (final concentration) of [4,5-³H]-leucine (Perkin Elmer United States, specific activity 160 Ci mmol⁻¹) as previously described for sediments (Frank et al., 2019). The samples were incubated in the dark for 4–8 h under in-situ temperatures. At the end of the incubation, the assimilation of leucine was halted by the addition of 100 µl of 100% trichloroacetic acid solution. The sediment samples were then sonicated (Symphony, VWR) for 10 min to remove the bacterial biomass from the sediments (Frank et al., 2017). The bacterial biomass extract was divided into three 1 ml aliquots and the micro-centrifugation technique was applied (Smith et al., 1992). Disintegration per minute (DPM) from each sample was read using a liquid scintillation counter (Packard Tri carb 2100). A conversion factor of 1.5 kg C mol⁻¹ with an isotope dilution factor of 2.0 was used to calculate the bacterial carbon assimilation rate (Simon and Azam, 1989). Blank samples (n = 3 in each sampling event) included sediments randomly collected from the benthocosms added with the [4,5-³H]-leucine and immediately with trichloroacetic acid, and incubated under the same conditions as described above.

DNA was extracted from the sediment samples (∼500 mg wet weight) with the PowerSoil DNA bead tubes (Qiagen, United States), using the FastPrep-24™ Classic (MP Biomedicals, United States) bead-beating to disrupt the cells (2 cycles at 5.5 m sec^–1, with a 5 min interval). The V4 region (∼ 300 bp) of the 16S rRNA gene was amplified from the DNA (∼50 ng) using the 515Fc/806Rc primers amended with CS1/CS2 tags (5′- ACACTGACGACATGGTTCTACAGTGYCAGCMGCCGCGGTAA, 5′- TACGGTAGCAGAGACTTGGTCTGGACTACNVGGGTWTCTAAT, (Apprill et al., 2015; Parada et al., 2016). PCR conditions were as follows: initial denaturation at 94°C for 45 s, 30 cycles of denaturation (94°C for 15 s), annealing (15 cycles at 50°C and 15 cycles at 60°C for 20 s) and extension (72°C for 30 s). Two annealing temperatures were used to account for the melting temperature of both forward (58.5–65.5°C), and reverse (46.9–54.5°C), primers. Library preparation from the PCR products and sequencing of 2 × 250 bp Illumina MiSeq reads was performed by HyLabs (Israel). The reads were deposited to the NCBI SRA archive under project number PRJNA882580.

Demultiplexed paired-end reads were processed in QIIME2 V2022.2 environment (Bolyen et al., 2019). Primers and sequences that did not contain primers were removed with cutadapt (Martin, 2011). Reads were truncated based on quality plots, checked for chimeras, merged and grouped into amplicon sequence variants (ASVs) with DADA2 (Callahan et al., 2016). The amplicons were classified with scikit-learn classifier that was trained on Silva database V138 (16S rRNA, Glöckner et al., 2017). Mitochondrial and chloroplast sequences were removed from the 16S rRNA amplicon dataset. Pathogen sequences were detected with 16SPIP (Miao et al., 2017).

Nutrients, cells abundance and activity were compared among the different hydrocarbon treatments using analysis of variance (ANOVA) followed by a Student-Newman-Keuls (SNK) post-hoc test (p < 0.05). Before analyses, the data were log-transformed to ensure the normality distribution of residuals (Shapiro-Wilk test) and checked for homogeneity of variance. The analyses were done by using the Microsoft add-in XLSTAT. We calculated beta diversity indices in R V3.6.3 (R Core Team 2021), using phyloseq (McMurdie and Holmes, 2013). Indispecies package was used to identify microbial lineages associated with treatment (Cáceres and Legendre, 2009).

Inorganic nutrient levels from the porewater exhibited large temporal variability in all treatments and controls (Figure 2). In the unamended control benthocosms, NO₂ ⁻ + NO₃ ⁻ ranged from 0.44–4.85 μmol L⁻¹ and PO₄ ³⁺ from 0.65–4.40 μmol L⁻¹, resulting in an average N:P ratio of ∼0.8:1. The NO₂ ⁻ + NO₃ ⁻ levels drastically declined in both the crude-oil and gas-condensate treatments in the first ∼100 days post hydrocarbon addition by up to 85%, and then gradually increased by 2–3 fold relative to the controls in the next ∼100 days (Figures 2A,B). Overall, the temporal alterations in NO₂ ⁻ + NO₃ ⁻ levels were similar between treatments statistically-wise, namely the control ≈ crude-oil ≈ gas-condensate (p> 0.05, ANOVA, Supplementary Table S1). PO₄ ³⁺ was rapidly consumed throughout the experimental duration in both treatments relative to the controls till it was nearly fully depleted (Figures 2C,D). Overall, the PO₄ ³⁺ levels in the control mesocosm > crude-oil ≈ gas-condensate (p< 0.05, ANOVA, Supplementary Table S1). The resulting N:P ratio was ∼2.8:1 in both the crude oil and the gas-condensate treatments. Additional chemical measurements directly related to the hydrocarbon additives in the top sediment included total organic carbon (TOC) and total polycyclic aromatic hydrocarbons (∑PAHs) concentrations that were measured at discrete time points along the experiment (days 0, 4, 74, and 207). The TOC level in the unamended control was ∼0.6% at all sampling points and increased to 1.17–1.32% in the crude oil and 0.76–1.02% in the gas-condensate benthocosms (Table 1). This resulted in ∼2-fold and ∼1.5-fold higher TOC values in the crude-oil and gas-condensate benthocosms, respectively, than the unamended controls. ∑PAHs levels were ∼18 ng gr⁻¹ in the control benthocosms and increased by as much as 6-fold higher concentrations in the hydrocarbon treatments (Table 1; Supplementary Table S2). Naphthalene and benzo(a)pyrene were the dominant species comprising the PAHs at all sampling points in both the crude-oil and gas-condensate treatments (29–43% and 26–35%, respectively) (Supplementary Table S2). Other molecules such as Benzo (g,h,i)perylene (4–12%), Phenanthrene (7–13%), Fluoranthene (0–12%), 2-Methynaphthalene (1–6%), Benz(a)anthracene (0–5%), Benzo(b)fluoranthene (0–3%), Dibenzofuran (0–3%), Chrysene (0–5%), and Indeno (1,2,3-cd)pyrene (0–7%) were also detectable in some time points\treatments but not in all, and were often below the limit of quantification (Supplementary Table S2). Overall, TOC in the control < crude oil ≈ gas condensate, while for ∑PAHs the levels in the control < gas-condensate < crude-oil (p< 0.05, ANOVA; Supplementary Table S2).

Heterotrophic bacterial abundances in sediments showed little and insignificant difference between treatments throughout the experiment with scattered temporal variability (Figure 3A; Supplementary Table S1). The bacterial cell number in the crude oil treatment ranged from ∼1 × 10⁵ cells gr⁻¹ to ∼12 × 10⁵ cells gr⁻¹, while in the gas-condensate treatment it ranged from ∼0.4 × 10⁵ cells gr⁻¹ to ∼10 × 10⁵ cells gr⁻¹. Similar abundances were recorded in the control benthocosms, with heterotrophic bacterial abundance ranging from ∼0.3 × 10⁵ to ∼9 × 10⁵ cells gr⁻¹ (Figure 3A). The resulting heterotrophic bacterial abundance in the crude oil or gas condensate treatments relative to the controls showed scattered temporal variability, except from days 137–210 where lower abundances were found (Figure 3B), following the increase in NO₂ ⁻ + NO₃ ⁻ (Figure 2B). Unlike the slight variations detected in heterotrophic bacterial abundances (Figures 3A,B), heterotrophic bacterial production rates (BP) almost immediately (within a few hours) and significantly increased post-crude oil or gas condensate additions relative to the controls (Figures 3C,D, Supplementary Table S1). BP ranged from 0.3–1.7 μg C gr⁻¹ h⁻¹ in the control benthocosms, whereas in the crude oil and gas condensate these ranged from 0.7–6.9 μg C gr⁻¹ h⁻¹ and 0.7–2.9 μg C gr⁻¹ h⁻¹, respectively (Figure 3C). The increase in BP rates relative to the unamended controls was highest in the crude oil (maximal change ∼1,600%) and lower in the gas condensate (maximal change ∼800%) treatments (Figure 3D). Overall, the BP rates in the control < gas condensate < crude oil (p< 0.05, ANOVA; Supplementary Table S1).

Benthic phytoplankton biomass derived from chlorophyll. a measurements showed little and insignificant differences between treatments, ranging from 0.002 to 0.02 µg gr⁻¹ (Figure 4A; Supplementary Table S1). Chlorophyll. a level in the gas condensate benthocosms was usually lower than the controls throughout the experiment duration (maximal decrease of ∼50%) (Figure 4B). In contrast, an increase in chlorophyll. a was recorded in the first few days post crude oil addition (∼50%) followed by a gradual decrease until the conclusion of the experiment (maximal decrease of ∼85%) (Figure 4B). Benthic cyanobacterial abundance ranges were in the same order of magnitude between treatments; control (1.4 × 10³–18 × 10³ cells gr⁻¹), crude oil (1.1 × 10³–45 × 10³ cells gr⁻¹) and gas condensate (1.2 × 10³–22 × 10³ cells gr⁻¹) treatments (Figure 4C), yet exhibited different temporal variability (Figure 4D; Supplementary Table S1). Crude oil addition lead to an increase in cyanobacteria during the first 40 days compared to the controls (maximal change ∼85%), followed by a continuous decrease throughout the remaining experimental period (Figure 4B). Gas condensate addition lead to an immediate decrease in cyanobacterial abundance (by ∼20%), which remained below the respective control benthocosms throughout most of the sampling points (Figure 4D). Picoeukaryotic phytoplankton abundance was low in all treatments; 0.003–0.017 cells gr⁻¹ in comparison to the controls (Figure 4E). The gas condensate addition lead to an overall decrease in picoeukaryotes throughout the whole experiment relative to the controls (maximal ∼80%), while crude oil amendments resulted in a rapid abundance decrease in the first 120 days post-addition (maximal ∼75%) followed by a moderate increase (maximal 40%) (Figure 4F; Supplementary Table S1).

Amplicon sequencing of the 16S rRNA gene showed that overall the benthic microbial communities were similar between treatments and the controls, complex and highly heterogeneous. Whereas alpha diversity appeared to increase in all the treatments compared to t0, no clear trends in alpha diversity were observed among the treatments, and there was ssubstantal noise in the data (Supplementary Figure S1). The dominant bacterial lineages were Alphaproteobacteria, Gammaproteobacteria, Bacteroidia and Planctomycetes (>10% of all ASVs each) (Supplementary Figure S2A). We observed drift of population structure in all treatments, in particular, at later stages, after a 48-h incubation (Supplementary Figure S2B). Despite the high heterogeneity of the samples, indicator species analysis revealed the enrichment of hydrocarbonoclastic bacteria following oil amendments (considering only the first two time points, before the drift in community composition) including Methylophaga, Ponticaulis and Alcanivorax genera (Table 2). Gas condensate amendments enriched a limited number of lineages whose metabolism is unknown as well as an Actinobacterota 67–14 lineage that has been associated with heavy metal contaminations in sediments (Table 2).

The ambient ∑PAHs level in the unamended (control) littoral sediments was higher (yet insignificantly, t-test, p = 0.057) than the values reported over the shelf (∼30–200 m) and slope (200–1,000 m) of the eastern Mediterranean Sea in pristine areas (Barakat et al., 2011; Kucuksezgin et al., 2013; Astrahan et al., 2017). This suggests that the littoral sediments used in the benthocosms were moderately polluted by hydrocarbons, to begin with, regardless of any experimental manipulation. The source of this chronic-low level pollution maybe be tracked to the 1970-80s as discussed above (Golik et al., 1988). Note, however, that the ∑PAHs values we measured, which represent current-nowadays conditions in the littoral zone, are below the corresponding Effects Range Low (ERL) standards (Long et al., 1995), and are consistent with other observations from the eastern Mediterranean (Mandalakis et al., 2014).

The TOC levels measured in this study are in agreement with previously reported TOC values from the shelf of the southeastern Mediterranean Sea (Astrahan et al., 2017). The variation in TOC values may be attributed to differences in grain sizes (de Haas et al., 2002) and other land-derived processes (Herut et al., 2000; Efrati et al., 2013). PAHs are often positively correlated with TOC in sediments (Lohmann, 2003; Oleszczuk and Baran, 2015). In the control benthocosms there was no clear relationship between TOC and ∑PAHs, in agreement with the different natural sources for these molecules (that is, PAHs are generally associated with anthropogenic pollution).

The ambient NO₂ ⁻ + NO₃ ⁻ and PO₄ ³⁺ concentrations in the porewater are in agreement with the low nutrient levels characteristic of this oligotrophic region (Sisma-Ventura et al., 2021). The corresponding average ∼1:1 N:P ratio of the inorganic fraction suggests N-limiting conditions for autotorphic microbial populations (Redfield, 1934; Moore et al., 2013). The ambient concentrations of hydrocarbons in eastern Mediterranean sediments (Golik et al., 1988) suggest that in the case of a hydrocarbon spill there may be a microbial inoculum of hydrocarbon degraders able to use and decompose some of the hydrocarbons. Our molecular results support this notion, showing the presence of hydrocarbon degraders such as Methylophaga and Alcanivorax in the control benthocosms, though in low abundances (Supplementary Figure S2). This is not surprising since hydrocarbon-degrading bacteria such as Alcanivorax were previously reported in the southeastern Mediterranean Sea, both in the sediments (Guy-Haim et al., 2020; Rubin-Blum et al., 2022) and in the water column (Keuter et al., 2015).

Our experiments simulated a moderate/high hydrocarbon spill in a coastal environment without physical mixing by wave movement. Under such conditions, most hydrocarbons likely bypass the 25 cm water layer as MOS and reach the sediments as sporadically-dispersed particles. This may explain the considerable heterogeneity in our results, as there likely was no consistent diffusion of hydrocarbons from the slick, but mainly formation and sinking of hydrocarbon-containing particles, that reach the sediments sporadically. Nonetheless, crude oil and gas condensate amendments resulted in a significant increase in ∑PAHs and TOC levels (Table 1). These enrichments were more pronounced in crude oil amendments, whereas gas condensate may have triggered a delayed response (i.e., the increase in ∑PAHs was observed only at the end time-point). A positive-linear relationship between TOC and ΣPAHs was observed (Pearson, p < 0.05), suggesting a common source or the existence of carbon moieties in the PAHs molecules. A similar positive relationship was also found in several stations on the southern part of the Israeli coast in proximity to gas drilling sites and production platforms (Astrahan et al., 2017), suggesting that in both cases TOC and PAHs are linked. However, the ∑PAHs concentration was below its ERL standards for coastal sediments (Long et al., 1995), and significantly lower than some chronically polluted sediments in the Mediterranean Sea such as Alexandria harbor (Barakat et al., 2002) and El-Arish or El-Mex (Aly Salem et al., 2013). This is in line with the moderate, yet prolonged, effects we observed in the incubations.

Our results indicate that both crude oil and gas condensate may trigger similarly prolonged (months and possibly more), yet usually moderate changes in benthic microbial abundances and diversity. This is surprising considering gas condensate consists of many volatile compounds and should potentially be less toxic on long timescales than crude oil (Øksenvåg et al., 2017). This suggests that the remaining non-volatile PAH molecules in gas condensate, either solely or cumulatively, impact benthic algae and bacteria. We also propose that volatile molecules from gas condensate attached to aerosols may be redeposited in the surface water downwind of the spill site with toxic effects. This issue should be examined in future studies.

The overall minor changes in heterotrophic microbial variables exposed to hydrocarbons may be explained in several not mutually exclusive ways: 1) Hydrocarbon degraders that were present before the amendments may become active at the expense of non-hydrocarbon degrading bacteria. Under these circumstances, bacterial abundance may not necessarily change, but selection of more tolerant microbes may occur, and bacterial activity is expected to increase; 2) Macronutrients such as N or P may limit bacterial growth post hydrocarbon additions (Shai et al., 2021) and 3) toxic effects of some PAH molecules (e.g., Naphthalene and\or benzo(a)pyrene, Supplementary Table S2) reduce bacterial growth variables (excluding hydrocarbon degraders). The decrease in NO₂ ⁻ + NO₃ ⁻ levels following crude oil or gas condensate additions relative to the controls in the first ∼100 days post-addition (Figures 2A,B and Supplementary Table S1), was likely due to increased bacterial use through assimilation and denitrification, accompanied by elevated bacterial production rates (Figures 3C,D and Supplementary Table S1). Indeed, crude oil amendments were previously shown to stimulate denitrification (Zhao et al., 2020). We may also consider the role of benthic dinitrogen (N₂) fixation, which could be responsible for the increase in NO₂ ⁻ + NO₃ ⁻ levels relative to the controls in the remaining days till the conclusion of the experiment (Figures 2A,B; Supplementary Table S1), as diazotrophs were found to tolerate oil-polluted sediments in an estuary system (Chronopoulou et al., 2013). Given the nitrogen limitation of hydrocarbon degradation (Musat et al., 2006), and the increased budget of energy-rich compounds that can support the costly N₂ fixation, diazotrophs may indeed have an advantage in the long term, producing nitrogen that can yield new NO₂ ⁻ + NO₃ ⁻. Notably, we did not measure denitrification or N₂ fixation rates or detect specific-related genes such as the nirK or nifH. Therefore, we cannot ascertain whether this interplay between different biochemical processes occurred. However, many denitrifying bacteria and heterotrophic diazotrophs comprise the bacterial classes we found using 16S rDNA analyses (Table 2, Supplementary Figure S2), so it is likely. Future investigation of physiological rates and microbial functions is needed to investigate the N budgets following hydrocarbon introduction to the sediments, either directly or through MOS.

Along with nitrogen, phosphorus is likely to be a co-limiting nutrient for hydrocarbon degraders in this area, exemplified by the increased bacterial activity coupled with the decline in PO₄ ³⁺. Unlike inorganic nitrogen, porewater PO₄ ³⁺ is rapidly consumed by benthic microorganisms, highlighting the P-limiting conditions for bacteria in this oligotrophic area (Sisma-Ventura and Rahav, 2019) and hence the commonly observed P influx from the water column to the sediment (Sisma-Ventura et al., 2021; Rubin-Blum et al., 2022). Given that orthophosphate has very few external and sporadic sources (e.g., dust deposition, Guieu et al., 2014), hydrocarbon-degrading bacteria in the eastern Mediterranean Sea are expected to face severe P-limitation. Contrary, nitrogen can be fixed biologically by benthic diazotrophs (Musat et al., 2006) therefore compensating for the low nitrogen availability.

Observations made in a temporal study following the Prestige oil spill off northwestern Spain (Alonso-Gutiérrez et al., 2009) indicated that bacterial abundances almost did not vary, following our results (Figures 3A,B, Supplementary Table S1). On the other hand, bacterial productivity increased significantly (Figures 3C,D), and bacterial diversity changed substantially (Supplementary Figure S1) in the treated benthocosms compared to the controls (Supplementary Table S1). One such alteration could be explained by the interplay between denitrifying bacteria and diazotrophs discussed above. Our results, therefore, suggest that monitoring bacterial activity is a powerful tool to pinpoint hydrocarbon pollution, whereas bacterial abundance measurements may be misleading or not indicative.

Given that PAHs, such as benzo(a)pyren and naphthalene that were prominent in our experiments (Supplementary Table S2) and are considered toxic to autotrophic microbes (Hylland, 2006; Ozhan and Bargu, 2014), the significant decrease in benthic picoeukaryote abundances following the addition of crude oil or gas condensate is not surprising. However, picoeukaryotes may be also outcompeted for nutrients (e.g., N, P) by the fast-growing hydrocarbon-degrading heterotopic bacteria. In addition, hydrocarbon-degrading bacteria may not only benefit from the increased availability of metabolic substrates but also “enjoy” the limited grazing pressure due to the toxic effects of the PAHs (Chronopoulou et al., 2013). Moreover, the increase in the ∑PAHs after the experiment in the gas condensate benthocosms (day 207, Table 2) may suggest increased toxicity to infaunal organisms, including microbes.

Alphaproteobacteria, Deltaproteobacteria and Gammaproteobacteria play a role in crude oil degradation due to their ability to use aliphatic and aromatic compounds (David et al., 2005; Joel et al., 2011; Doyle et al., 2018). Bacteroidetes that can also use hydrocarbons occur in oil-contaminated sediments (Paissé et al., 2008; Sun et al., 2013; Korlević et al., 2015) (Yakimov et al., 2007; Quero et al., 2015). At higher taxonomic ranks, hydrocarbon-degrading bacteria were represented by Oceanospirillales, Cellvibrionales, Alteromonadales, Sphingomonadales, Steroidobacterales, and Flavobacteriales orders, and particularly by Methylophaga and Alcanivorax genera, in both the hydrocarbon-added benthocosms and the controls. This suggests that the benthic microbial communities in our study site routinely encounter hydrocarbons, as discussed above.