The application of dispersants to an oil-slick is a key remediation tool and thus understanding its effectiveness is vital. Two in situ oil slicks were created in the North Sea (off the coast of The Netherlands), one left to natural processes whilst dispersant (Slickgone NS) was applied to the other. GC-MS analysis of seawater from the surface slick, and at 1.5 and 5 m below the slick, revealed only two samples with measurable hydrocarbons (221 ± 92 μg ml−1 seawater), from the surface of the “Slickgone Dispersed” oil-slick ~25.5 hours after oil-slick formation, which was likely due to environmental conditions hindering sampling. Additionally, 16S rRNA gene quantitative PCR and amplicon analysis revealed extremely limited growth of obligate hydrocarbonoclastic bacteria (OHCB), detected at a relative abundance of <1×10-6 %. Furthermore, the Ecological Index of Hydrocarbon Exposure (EIHE) score, which quantifies the proportion of the bacterial community with hydrocarbon-biodegradation potential, was extremely low at 0.012 (scale of 0 – 1). This very low abundance of hydrocarbon-degrading bacteria at the time of sampling, even in samples with measurable hydrocarbons, could potentially be attributed to nutrient limitation (~25.5 hours after oil-slick creation total inorganic nitrogen was 3.33 μM and phosphorus was undetectable). The results of this study highlight a limited capacity for the environment, during this relatively short period, to naturally attenuate oil.

The overall goal of oil-spill response is to minimise impact to life as well as natural and economic resources. A balance must be made between potential environmental/economic impacts and “natural recovery” or “recovery through intervention” (National Oceanic and Atmospheric Administration, 2010; IPIECA et al., 2017). Oil-spill response in the marine environment requires a comprehensive knowledge of immediate and surrounding environments, local stakeholders and political legislation, and available remediation tools. One such tool is the application of dispersants. Dispersants transform oil on the surface of the water into droplets (10 – 300 μm, North et al., 2015) in the water column, which increases oil surface area for microbial attachment (Prince et al., 2013), thus allowing hydrocarbon-degrading microbes to expend more energy on growth and less energy producing biosurfactants, thereby expediting hydrocarbon-biodegradation (Prince et al., 2016; Brakstad et al., 2018). Prince et al. (2015) found that three commonly applied dispersants (Corexit 9500, Finasol OSR 52, and Slickgone NS) significantly increased the biodegradation of hydrocarbons when compared to a floating oil-slick, with no added dispersants. Several other studies also observe that dispersants increase biodegradation (Brakstad et al., 2015; Prince et al., 2016) and enhance the growth of hydrocarbon-degrading bacteria (HCB) (Hazen et al., 2010; Dubinsky et al., 2013; Ribicic et al., 2018). In contrast, other studies show that dispersants may not enhance biodegradation (Lindstrom and Braddock, 2002; Rahsepar et al., 2016) or may even inhibit the growth of HCB (Hamdan and Fulmer, 2011; Kleindienst et al., 2015); though there are many criticisms of the experimental procedures used in these studies (Gregson et al., 2021).

With such contradictory results from studies investigating the effects of dispersant application on oil spills it is evident that further research is required. Studies must replicate natural environmental conditions as best they can. The optimal way would be to collect samples during the application of a dispersant to a real oil-spill. However, due to logistical, financial, and safety issues this is often not possible. The next best option is to conduct a controlled experiment in situ. However, once again there are many legislative, economic, and technical barriers to such experiments being approved. Due to limitations sampling real and experimental oil spills, most oil/dispersant research is conducted in the laboratory (Buist et al., 2011; Tremblay et al., 2017; Doyle et al., 2018). However, conducting ex-situ oil-spill experiments has potentially negative biases due to confinement in bottles or tanks, which would not necessarily occur at sea. Confinement does not allow dispersed oil to rapidly dilute to sub-ppm concentration which would occur in situ (Bejarano et al., 2013), and higher concentrations of oil can potentially inhibit hydrocarbon degradation (Lee et al., 2013; Prince et al., 2016). Moreover, dispersion over a wider area, which would occur at sea, may allow access to further inorganic nutrients, which in turn could lead to faster hydrocarbon degradation. Given the potential biases of conducting laboratory oil-spill experiments, obtaining a permit to conduct a controlled oil release at sea, with dispersant application, is highly valuable.

The North Sea, in the north eastern area of the Atlantic Ocean, is located between the United Kingdom and borders continental west Europe. Approximately 13 miles off the coast of Scheveningen harbor, The Hague, Netherlands, a controlled in situ experiment was conducted in which oil slicks were created in April 2019. One oil-slick was left to undergo natural attenuation and dispersion whilst the other oil slick was chemically dispersed using the widely applied commercial dispersant Slickgone NS (Dasic International). Samples were taken from both oil slicks approximately 1, 5.5, and 25.5 hours after oil-slick creation, providing a rare and valuable opportunity to evaluate whether dispersant application on oil spills affects HCB growth and hydrocarbon-biodegradation in situ.

Sampling Campaign: Oil-slick creation took place on the 16th April 2019 approximately 13 miles off the coast of Scheveningen harbour, The Hague, Netherlands. Full sampling and technical details can be found in the ITOPF ExpOS'D technical report (Zeinstra et al., 2020), but in summary, a light-medium Arabian Crude oil was released continuously, using an air membrane pump with a flow rate of 6.7 litres s−1, via a 2-inch hose. This trailed 20 m behind the vessel, on floatation bladders, travelling at 1.85 knots. The natural dispersion oil slick (“Naturally Dispersed”) was created into the wind at 10:25, using ~2.5 m3 of the crude oil. The oil slick for chemical dispersion (“Slickgone Dispersed”) was created into the wind at 11:40, using ~2.5 m3 of the crude oil. The “Slickgone Dispersed” oil slick was sprayed (by an onboard MARKLEEN Dispersant spray system) with Slickgone NS dispersant 30 minutes after release, for one hour at a ratio of 20:1 oil to dispersant. Triplicate 250 ml seawater samples were taken from the surface and at depths of 1.5 m and 5 m in sterile plastic containers. Surface samples were taken directly by reaching over the side of the rigid inflatable boat (RIB). Samples from 1.5 and 5 m depths were taken by means of a sterile hose, lowered to the required depth, and samples pumped into sterile plastic containers. Sample water (150 ml) was passed through Millipore® Sterivex™ filters (0.22 μm) and flash frozen at −150°C in a Cryogenic Vapour Shipper, to preserve DNA, prior to storage at −20°C. The filtrate from this process was also flash frozen prior to being stored at −20°C for nutrient analysis of ammonium (NH4+), phosphate (PO43−), nitrate (NO3), and nitrite (NO2), using a SEAL Analytical AA3 HR AutoAnalyzer tandem JASCO FP-2020 Plus fluorescence detector. In addition, triplicate 40 ml seawater samples were collected from the surface as well as at depths of 1.5 m and 5 m (same method as above), in sterile brown-glass 40 ml vials capped with PTFE-lined silicon septa, and immediately frozen at −20°C for hydrocarbon analysis. Sampling of oil slicks occurred ~1.5 hours, ~5 hours (16th April 2019), and ~25.5 hours (17th April 2019) after oil-slick creation.

Environmental Measurements: Temperature (9.06 ± 0.11 °C), salinity (30.9 ± 0.85 psu), and pH (8.41 ± 0.02) were all measured at the time of sampling. Wave height measurements were collected by two stations: ‘IJgeul 1' (4,264°E, 52,488°N, located 31 km of sampling site) and ‘Q1 platform' (4,150°E, 52,925°N, located 75 km northeast of sampling site). Wind speed/direction measurements were collected by two offshore stations: P11 (3,342°E, 52,359°N, 45 km northwest of sample site) and Europlatform (3,275°E, 51,998°N, 55 km southwest of sample site) (Zeinstra et al., 2020).

Hydrocarbon Degradation (GC-MS): Hydrocarbons were extracted from 40 ml brown-glass vials (collected in situ) using a 20 ml solvent extraction of 1:1 hexane : dichloromethane, vigorously shaken for 30 seconds, and placed in an ultrasonic bath for 30 minutes. The 20 ml of solvent extract was then passed through reversed-phase solid-phase extraction tubes (Supelclean ENVI-18 SPE, Sigma), using an method adapted from Risdon et al. (2008), before being eluted in 6 ml of 1:1 hexane : dichloromethane and then concentrated to 1 ml under nitrogen gas. Sample quantification was performed on an Agilent 7890A Gas Chromatography system coupled with a Turbomass Gold Mass Spectrometer with Triple-Axis detector, operating at 70 eV in positive ion mode, using conditions as previously described by Coulon et al. (2007). Only those hydrocarbons detected are shown in Fig. 1 (B).

Fig. 1:

profile of the light-medium Arabian Crude deposited during oil-slick formation; including n-alkanes (C11 to C31), branched alkanes pristane and phytane, and PAHs (naphthalene, fluorene, phenanthrene, and any methylated derivatives (naphthalene and phenanthrene/anthracene (Phen/Anth)) (A). Concentration of measured hydrocarbons from seawater samples taken from the “Slickgone Dispersed” oil-slick, ~25.5 hours after oil-slick creation (B).

Fig. 1:

profile of the light-medium Arabian Crude deposited during oil-slick formation; including n-alkanes (C11 to C31), branched alkanes pristane and phytane, and PAHs (naphthalene, fluorene, phenanthrene, and any methylated derivatives (naphthalene and phenanthrene/anthracene (Phen/Anth)) (A). Concentration of measured hydrocarbons from seawater samples taken from the “Slickgone Dispersed” oil-slick, ~25.5 hours after oil-slick creation (B).

Close modal

qPCR Analysis of Bacterial 16S rRNA genes: DNA was extracted from in situ seawater samples from thawed Millipore® Sterivex filters with a DNeasy PowerWater Sterivex Kit (Qiagen) according to the manufacturer's instructions. The primers used for quantification of bacterial 16S rRNA genes were 341f - CCTACGGGNGGCWGCAG and 785r – GACTACHVGGGTATCTAATCC (Klindworth et al., 2013). qPCR was performed using a CFX384™ Real-Time PCR Detection System (BioRad) using reagents, cycle conditions, and standards as previously described (McKew and Smith, 2015).

Amplicon Sequencing and Bioinformatics: Amplicon libraries were prepared, as per Illumina instructions. PCR primers were the same as those used for qPCR but flanked with Illumina overhang sequences. PCR products were quantified using Quant-iT PicoGreen dsDNA Assay Kit (ThermoFisher Scientific) and pooled in equimolar concentrations. Quantification of the amplicon libraries was determined via NEBNext® Library Quant Kit for Illumina (New England BioLabs Inc.), prior to sequencing on the Illumina MiSeq® platform, using a MiSeq® 600 cycle v3 reagent kit and 20% PhiX sequencing control standard. Sequence output from the Illumina MiSeq platform were analysed within BioLinux (Field et al., 2006), using a bioinformatics pipeline as described by Dumbrell et al. (2016).

Statistical Analysis: Prior to community analysis, sequence data were rarefied to the lowest library sequence value (5,747). Data were first tested for normality (Shapiro-Wilks test), those data which were normally distributed were tested for significance with ANOVAs or appropriate linear models. Non-normally distributed data were analysed using appropriate GLMs (Generalised Linear Models) as follows. The relative abundance of operational taxonomic units (OTUs) or genera in relation “Uncontaminated Seawater”, both oil-slicks, depth, or time were modelled using multivariate negative binomial GLMs (Wang et al., 2010). Here, the number of sequences in each library was accounted for using an offset term, as described previously (Alzarhani et al., 2019). The abundance of bacterial 16S rRNA gene copies was also modelled using negative binomial GLMs (Venables and Ripley, 2002). The significance of model terms was assessed via likelihood ratio tests. The Environmental Index of Hydrocarbon Exposure (Lozada et al., 2014) was calculated using the script available at the ecolFudge GitHub page (https://github.com/Dave-Clark/ecolFudge, Clark, 2019) and EIHE values modelled using poisson GLMs. All statistical analyses were carried out in R3.6.1 (R Development Core Team, 2011) using a variety of packages available through the references (Venables and Ripley, 2002; Csardi and Nepusz, 2006; Hope, 2013; Wilke, 2015, 2020; Becker et al., 2016; Auguie, 2017; Oksanen et al., 2019; Hvitfeldt, 2020; Kassambara, 2020; Lenth, 2020; Pedersen, 2020). All plots were constructed using the “ggplot2” (Bodenhofer et al., 2011) and “patchwork” (Pedersen, 2019) R packages.

Hydrocarbon Analysis Reveals Difficulty in Conducting in situ Oil-spill Experiments

Analysis of hydrocarbons revealed that only two samples, from the “Slickgone Dispersed” oil-slick ~25.5 hours after oil-slick creation, contained any measurable hydrocarbons; including n-alkanes (C14 – C31), branched alkanes (pristane and phytane), and polycyclic aromatic hydrocarbons (PAHs; phenanthrene and methyl-phenanthrene/anthracene) at average concentrations of 188.13 (± 76.91), 27.20 (± 11.91), and 5.84 (± 3.13) μg ml−1 seawater, respectively (Fig. 1B). These samples did not contain any measurable C11 – C13n alkanes or naphthalenes and fluorene, in comparison to a profile of the oil (Fig. 1A), suggesting these hydrocarbons has partitioned into the air and/or water. Furthermore, the ratio of n-C17/pristane and n-C18/phytane was 0.95 and 1.63, respectively, with no significant difference to the original oil (n-C17/pristane (0.94) and n-C18/phytane (1.47)), indicating no biodegradation. A similar in situ North Sea oil spill by Gros et al. (2014) observed rapid mass transfer of >50% of <C17 hydrocarbons, as well as no detectable naphthalene, from surface samples 25 hours after oil-slick creation. The lack of measurable hydrocarbons, in all other surface samples from this field trial, was despite the fact oil was clearly visible to the naked eye and via radar, at all sampling time points. Samples were taken by reaching out of a rigid inflatable boat (RIB) and collecting surface oil/water in sterile vials. However, this proved difficult during the first day (16.04.2019) as increased wind speeds and wave heights, 8.33 ± 0.71 m s−1 and 105.26 ± 17.32 cm respectively, bounced the RIB, pushing the oily surface water beyond reach. On the second day, wind speed and wave height reduced to 5.15 ± 0.66 m s−1 and 58.52 ± 7.63 cm respectively, and samples were collected by means of a vial attached to a 2 m stick. Whilst the calmer environmental conditions and the new sampling technique meant sampling the oil/water interface was easier, movement of the RIB still made it difficult, resulting in only 2 of the 9 surface samples, collected ~25.5 hours after oil-slick creation, having any measurable hydrocarbons. These results reflect the difficulty in efficiently obtaining in situ oil-spill samples from the surface oil/water interface. Samples collected at depths of 1.5 and 5 m would not have been affected as seawater was directly pumped from those depths into sterile vials, suggesting oil either remained on the surface or had dispersed beyond these depths.

The overarching criticism of ex situ oil-spill experiments is that the oil spills are enclosed by some form of container, be it a microcosm, mesocosm, or wave tank. This containment is believed to create a number of biases, one of which is that containment decreases oil dispersal and dilution, which would otherwise dilute to sub-ppm concentrations in situ within 1 to 4 hours (Nedwed and Coolbaugh, 2008; Bejarano et al., 2013). Therefore, adding oil at greater concentrations than sub-ppm, may inhibit the growth of some hydrocarbon-degrading bacteria (HCB), and thus reduce the rate of hydrocarbon biodegradation (Prince, et al., 2016). The results of this study could suggest that the concentration of oil from marine oil slicks that have been sprayed with dispersant does not always reduce to sub-ppm immediately, as the two samples with measurable oil (from the surface of the “Slickgone Dispersed” oil-slick ~25.5 hours after oil-slick creation) contained hydrocarbons at ~221 ppm. Moreover, the “Slickgone Dispersed” oil-slick, whilst reduced in size, remained visible by radar at all time-points. The application of dispersants to an oil slick requires suitable environmental conditions, which include wind speeds of 4 – 12 m s−1 (ITOPF, 2011) and full salinity seawater at 32–35 psu (Chandrasekar et al., 2006). Additionally, dispersant efficacy is affected by the type of oil, as increasing oil viscosity decreases dispersant effectiveness, and therefore its application is more suited to light-to-medium oils (Trudel et al., 2010). Weathering of oil increases viscosity, and thus the window of opportunity to apply dispersants to oil slicks ranges from a few hours to a few days (Chandrasekar et al., 2005; ITOPF, 2011). These criteria were met during this study, and therefore it is unlikely that environmental conditions (wind speed, wave height, and salinity), oil type (light-medium Arabian Crude), or window of opportunity (one hour after oil-slick creation), inhibited dispersant efficiency. It should be noted, however, that the application of Slickgone NS on the oil slick was below the recommended level to sufficiently coat the oil-slick. Approximately 200 litres of dispersant was applied to the oil-slick, however, this is considerably lower than the 700 litres required to achieve the manufacturer's recommendation of 40 – 50 L per 10,000 m2 of oiled area (Zeinstra et al., 2020). This was due to time constraints restricting the number of dispersant-spraying passes through the oil-slick, thus not all areas of the slick had dispersant applied. None of the samples pumped directly from 1.5 or 5 m depths contained any measurable hydrocarbons, suggesting that either where the seawater was sampled the dispersant had not been applied to that part of the oil-slick, or that, had the dispersant been applied to that area, the oil had already been dispersed beyond 5 m.

Nutrient Limitation Potentially Inhibited the Growth of Hydrocarbon-degrading Bacteria

Certain microbes can degrade a range of hydrocarbons found in crude oil and its derivatives and thus oil-spills dramatically alter marine microbial community composition, resulting in a decrease in species richness and diversity, in conjunction with selection for HCB (Head et al., 2006; McGenity et al., 2012). However, during this study there was a clear lack of growth of OHCB or those genera with known hydrocarbon-degrading species. The Ecological Index of Hydrocarbon Exposure (EIHE), which quantifies the proportion of the bacterial community with hydrocarbon-biodegradation potential (Lozada et al., 2014), was extremely low, averaging 0.012 (± 0.003; scale of 0 – 1) over all samples.

There were no significant differences in the EIHE score between “Uncontaminated Seawater” and each oil-slick at all time points and depths (Fig. 2A–C). Compared to some other marine environments (Fig. 2D) it can be observed that the EIHE score of 0.012 is similar to that found in sediments around the WWII shipwreck HMS Royal Oak, where the EIHE score was 0.008 and PAH levels were 229.2 ± 126.5 μg kg−1 of dry sediment (Thomas et al., unpublished). Furthermore, the EIHE score of 0.012 observed in this study is much lower than the EIHE scores observed in contaminated sediments (EIHE 0.52; TPH 1,093 – 3,773 μg g−1 dry sediment) sampled five-days after the Agia Zoni II oil-spill (Thomas et al., 2020) and in oil/dispersant North Sea seawater samples taken after 24 hours (EIHE 0.50; TPH 54.95 μg ml−1) from an oil/dispersant microcosm experiment (Thomas et al., unpublished). The relative abundance (%) of genera assigned to obligate hydrocarbonclastic bacteria (OHCB), a group of widely distributed marine bacteria that are specifically adapted to using hydrocarbons as an almost exclusive source of carbon and energy (Yakimov et al., 2007), was less than 1×10−6.

Fig. 2:

Ecological Index of Hydrocarbon Exposure (EIHE) scores (± SE, n = 3, ratio % up to 1), representing relative abundance of bacteria with hydrocarbon-biodegradation potential (Lozada et al., 2014), from seawater sampled over ~25.5 hours (A), and over a 5 m depth profile (B), from “Uncontaminated Seawater” and “Naturally Dispersed” and “Slickgone Dispersed” oil-slicks (C). Additionally, a comparison between EIHE scores from seawater samples taken in this study (“North Sea Oil-slicks”) and other marine environments (D): “HMS Royal Oak” (Thomas et al., unpublished, average over all samples), “North Sea Oil-slicks” (this study), “North Sea Oiled Microcosms” (Thomas et al., unpublished; average over dispersant treatments (which reduced oil/water interfacial tension) after 24 hours), and “Agia Zoni II Oil-spill” (Thomas et al., 2020, average at impacted sites in September 2017).

Fig. 2:

Ecological Index of Hydrocarbon Exposure (EIHE) scores (± SE, n = 3, ratio % up to 1), representing relative abundance of bacteria with hydrocarbon-biodegradation potential (Lozada et al., 2014), from seawater sampled over ~25.5 hours (A), and over a 5 m depth profile (B), from “Uncontaminated Seawater” and “Naturally Dispersed” and “Slickgone Dispersed” oil-slicks (C). Additionally, a comparison between EIHE scores from seawater samples taken in this study (“North Sea Oil-slicks”) and other marine environments (D): “HMS Royal Oak” (Thomas et al., unpublished, average over all samples), “North Sea Oil-slicks” (this study), “North Sea Oiled Microcosms” (Thomas et al., unpublished; average over dispersant treatments (which reduced oil/water interfacial tension) after 24 hours), and “Agia Zoni II Oil-spill” (Thomas et al., 2020, average at impacted sites in September 2017).

Close modal

Potentially, the growth of HCB was inhibited by the absence of nutrients, where the level of total inorganic nitrogen (TIN; sum of ammonia, nitrate, and nitrite) significantly decreased in all samples (23.74 to 3.33 μM) ~25.5 hours after oil-slick creation, as well as phosphate being undetected (Fig. 3); the limit of detection for nutrients was 0.02 μM. Both nitrogen (N) and phosphorous (P) are vital for microbial growth, for example, N is required for the synthesis of proteins and nitrogenous bases whilst P is required for the synthesis of nucleic acids and phospholipids (Bristow et al., 2017). N and P are especially important during hydrocarbon degradation of an oil slick (Atlas, 1981), and therefore the availability of these nutrients in the presence of hydrocarbons is vital (Ron and Rosenberg, 2014). Certain HCB, such Alcanivorax and Cycloclasticus, have specific systems for scavenging nutrients in oligotrophic environments (Wang et al., 1996; Cappello and Yakimov, 2010). However, the lack of growth of microbes such as Alcanivorax and Cycloclasticus species and other OHCB, suggests P limitation, or that growth was limited in some other way. The concentration of nutrients in the North Sea is primarily driven by a seasonal cycle, with higher levels of N and P in the winter months compared to the summer months (Tett and Walne, 1995). It is likely that the rapid decline of TIN, over ~1 day, was due to a decrease in vertical mixing as wave energy declined. Phytoplankton blooms, which take place during times of increased sunlight and nutrients in the euphotic zone, often occur in the spring and last until summer when nutrients become depleted (Mann and Lazier, 2013). Satellite images captured by MODIS (Moderate Resolution Imaging Spectroradiometer) suggest a phytoplankton bloom in the North Sea began on March 29th, 2019 (NASA, 2019). Sampling of this study occurred on the 16th and 17th April 2019 and therefore the high abundance of phytoplankton could have depleted phosphorous.

Fig. 3:

Nitrate (A), ammonium (B), and nitrite (C) (mean ± SE, n = 3) from seawater samples taken from “Uncontaminated Seawater” as well as “Naturally Dispersed” and “Slickgone Dispersed” oil-slicks, ~1.5, ~5, and ~ 25.5 hours after oil-slicks were created. Phosphate was undetected in all samples; the limit of detection was 0.02 μM.

Fig. 3:

Nitrate (A), ammonium (B), and nitrite (C) (mean ± SE, n = 3) from seawater samples taken from “Uncontaminated Seawater” as well as “Naturally Dispersed” and “Slickgone Dispersed” oil-slicks, ~1.5, ~5, and ~ 25.5 hours after oil-slicks were created. Phosphate was undetected in all samples; the limit of detection was 0.02 μM.

Close modal

An EIHE score of 0.012 in this study reveals an exceptionally low level of HCB within seawater samples, this includes the two samples which contained hydrocarbons; though the microbial community samples were not truly paired to the samples collected for hydrocarbon analysis as they were collected in separate bottles (although at the same time and area of the slick) and therefore may not have contained oil. However, this demonstrates, at the time of sampling, the environment's ability to naturally attenuate oil was limited. Potentially low levels of phosphorous limited HCB growth, but it cannot be said for certain that this was the limiting factor. Regardless of what is limiting the growth of HCB, such low levels can inform oil-spill response operations. In this short-term study a limited ability for the environment to naturally attenuate oil would highlight a requirement for intervention measures, such as dispersal or physical removal of oil.

Developing in situ Experimental Oil-spill Methodologies

The results of this study have highlighted challenges in obtaining meaningful and reproducible seawater surface samples that capture the oil/water interface, with only 2 of the 18 surface samples from the oil-slicks containing any measurable hydrocarbons. The two samples with measurable hydrocarbons were taken from the “Slickgone Dispersed” oil-slick ~25.5 hours after oil-slick creation, though the third of the replicates contained no measurable hydrocarbons. Moreover, microbial community samples are not truly paired to the hydrocarbon samples as these were collected in separate vials for either DNA or hydrocarbon extraction. The primary challenge was the collection of surface samples from the sampling RIB, which would push the oily surface water beyond reach, even in relatively calm waters. One potential solution could be to use a remotely operated surface vehicle (ROSV), which could be remotely piloted (or done autonomously via GPS way-points) into the oil slick with minimal disturbance, collect a surface sample before returning to the crew for downstream processing. A ROSV designed and built for the purpose of oil-spill detection and sampling (e.g. Al Maawali et al., 2019) could be adapted further. Given a large enough capacity, the ROSV could even be adapted to apply dispersant at a specific location, which could then immediately be sampled, avoiding any doubt as to the efficiency of dispersant application. The efficacy of dispersant application was another limitation observed during this in situ oil-spill trial. This was primarily driven by time constraints, resulting in only 200 L, of the recommended 700 L, of dispersant actually being applied to the oil-slick. Technical recommendations advise more spraying passes through the oil-slick and that the dispersant spraying arms should be attached as far to the front of the ship as possible to ensure contact with oil, before it is pushed away by the ship's bow (Zeinstra et al., 2020). Ideally more time would be allocated in such trials, which would allow sufficient and effective dispersant application and sampling to occur. Moreover, longer field trials would allow biodegradation to be measured over more realistic timescales. Whilst rapid growth of hydrocarbon-degrading bacteria can be observed within laboratory seawater-oil microcosms within 24 hours (Thomas et al., unpublished), there is limited evidence that there would be significant growth in situ in many open water environments, and degradation would typically be very limited in the first day, particularly in low nutrient systems where a significant lag phase may be observed. However, due to permit restrictions requiring all oil to be removed from the sea surface after one day, high financial costs of operating numerous research vessels, and the availability of supporting services (i.e. airborne surveillance), additional time is not always possible. It is crucial any sampling limitations are overcome, as in situ oil-spill experiments can provide insightful results and observations into the processes that drive the fate and transport of oil in marine waters and thus guide oil-spill response management.

We would like to acknowledge and thank Rijkswaterstraat for facilitating the ExpOS'D project. We would also like to acknowledge and thank ITOPF for funding of this work (R&D Award 2018, ExpOS'D). Additionally, we would like to acknowledge and thank the National Environmental Research Council (NERC) (NE/L002582/1), via EnvEast DTP in CASE partnership with the Centre for Environment, Fisheries and Aquaculture Science, for funding this work. Further details about this study and other projects that formed the ExpOS'D (Experimental Oil Spill Data-sharing) project can be viewed in the ITOPF technical report at https://www.itopf.org/fileadmin/data/Documents/RDaward/ExpOS_D_Final_Report.pdf.

Alzarhani,
A.K.,
Clark,
D.R.,
Underwood,
G.J.C.,
Ford,
H.,
Cotton,
T.E.A.,
and
Dumbrell,
A.J.
(
2019
)
Are drivers of root-associated fungal community structure context specific?
ISME J
13
:
1330
1344
.
Atlas,
R.M.
(
1981
)
Microbial degradation of petroleum hydrocarbons: An environmental perspective
.
Microbiol Rev
45
:
180
209
.
Auguie,
B.
(
2017
)
gridExtra: functions in Grid graphics. R Package Version 2.3
.
CRAN Proj
.
Becker,
R.A.,
Wilks,
A.R.,
Brownrigg,
R.,
Minka,
T.P.,
and
Deckmyn,
A.
(
2016
)
Package “maps”: Draw Geographical Maps
.
R Packag version 23-6
.
Bejarano,
A.C.,
Levine,
E.,
and
Mearns,
A.J.
(
2013
)
Effectiveness and potential ecological effects of offshore surface dispersant use during the Deepwater Horizon oil spill: A retrospective analysis of monitoring data
.
Environ Monit Assess
185
:
10281
10295
.
Bodenhofer,
U.,
Kothmeier,
A.,
and
Hochreiter,
S.
(
2011
)
Apcluster: An R package for affinity propagation clustering
.
Bioinformatics
27
:
2463
2464
.
Brakstad,
O.G.,
Nordtug,
T.,
and
Throne-Holst,
M.
(
2015
)
Biodegradation of dispersed Macondo oil in seawater at low temperature and different oil droplet sizes
.
Mar Pollut Bull
93
:
144
152
.
Brakstad,
O.G.,
Ribicic,
D.,
Winkler,
A.,
and
Netzer,
R.
(
2018
)
Biodegradation of dispersed oil in seawater is not inhibited by a commercial oil spill dispersant
.
Mar Pollut Bull
129
:
555
561
.
Bristow,
L.A.,
Mohr,
W.,
Ahmerkamp,
S.,
and
Kuypers,
M.M.M.
(
2017
)
Nutrients that limit growth in the ocean
.
Curr Biol
27
:
R474
R478
.
Buist,
I.,
Potter,
S.,
Nedwed,
T.,
and
Mullin,
J.
(
2011
)
Herding surfactants to contract and thicken oil spills in pack ice for in situ burning
.
Cold Reg Sci Technol
67
:
3
23
.
Cappello,
S.
and
Yakimov,
M.
(
2010
)
Alcanivorax.
In
Handbook of Hydrocarbon and Lipid Microbiology
.
McGenity,
T.J.
(
ed
).
Berlin Heidelberg
:
Springer-Verlag
,
pp.
1737
1748
.
Chandrasekar,
S.,
Sorial,
G.A.,
and
Weaver,
J.W.
(
2006
)
Dispersant effectiveness on oil spills - impact of salinity
.
ICES J Mar Sci
63
:
1418
1430
.
Chandrasekar,
S.,
Sorial,
G.A.,
and
Weaver,
J.W.
(
2005
)
Dispersant effectiveness on three oils under various simulated environmental conditions
.
Environ Eng Sci
22
:
324
336
.
Clark,
D.R.
(
2019
)
ecolFudge.
Coulon,
F.,
McKew,
B.A.,
Osborn,
A.M.,
McGenity,
T.J.,
and
Timmis,
K.N.
(
2007
)
Effects of temperature and biostimulation on oil-degrading microbial communities in temperate estuarine waters
.
Environ Microbiol
9
:
177
186
.
Csardi,
G.
and
Nepusz,
T.
(
2006
)
The igraph software package for complex network research
.
InterJournal Complex Syst
.
Doyle,
S.M.,
Whitaker,
E.A.,
De Pascuale,
V.,
Wade,
T.L.,
Knap,
A.H.,
Santschi,
P.H.,
et al. (
2018
)
Rapid formation of microbe-oil aggregates and changes in community composition in coastal surface water following exposure to oil and the dispersant corexit
.
Front Microbiol
9
:
1
16
.
Dubinsky,
E.A.,
Conrad,
M.E.,
Chakraborty,
R.,
Bill,
M.,
Borglin,
S.E.,
Hollibaugh,
J.T.,
et al. (
2013
)
Succession of hydrocarbon-degrading bacteria in the aftermath of the deepwater horizon oil spill in the gulf of Mexico
.
Environ Sci Technol
47
:
10860
10867
.
Dumbrell,
A.J.,
Ferguson,
R.M.W.,
and
Clark,
D.R.
(
2016
)
Microbial Community Analysis by Single-Amplicon High-Throughput Next Generation Sequencing: Data Analysis – From Raw Output to Ecology
.
In
Hydrocarbon and Lipid Microbiology Protocols
.
McGenity,
T.J.,
Timmis,
K.N.,
and
Nogales,
B.
(
eds
).
Berlin, Heidelberg, Heidelberg
:
Springer Protocols Handbooks
,
pp.
155
206
.
Field,
D.,
Tiwari,
B.,
Booth,
T.,
Houten,
S.,
Swan,
D.,
Bertrand,
N.,
and
Thurston,
M.
(
2006
)
Open software for biologists: From famine to feast
.
Nat Biotechnol
24
:
801
803
.
Gregson,
B.H.,
McKew,
B.A.,
Holland,
R.D.,
Nedwed,
T.J.,
Prince,
R.C.,
and
McGenity,
T.J.
(
2021
) (
In Press
)
Marine Oil Snow, a Microbial Perspective
.
Front Mar Sci
.
Gros,
J.,
Nabi,
D.,
Würz,
B.,
Wick,
L.Y.,
Brussaard,
C.P.D.,
Huisman,
J.,
et al. (
2014
)
First day of an oil spill on the open sea: Early mass transfers of hydrocarbons to air and water
.
Environ Sci Technol
48
:
9400
9411
.
Hamdan,
L.J.
and
Fulmer,
P.A.
(
2011
)
Effects of COREXIT® EC9500A on bacteria from a beach oiled by the Deepwater Horizon spill
.
Aquat Microb Ecol
63
:
101
109
.
Hazen,
T.C.,
Dubinsky,
E.A.,
DeSantis,
T.Z.,
Andersen,
G.L.,
Piceno,
Y.M.,
Singh,
N.,
et al. (
2010
)
Deep-sea oil plume enriches indigenous oil-degrading bacteria
.
Science (80- )
330
:
204
208
.
Head,
I.M.,
Jones,
D.M.,
and
Röling,
W.F.M.
(
2006
)
Marine microorganisms make a meal of oil
.
Nat Rev Microbiol
4
:
173
182
.
Hope,
R.M.
(
2013
)
Rmisc: Ryan Miscellaneous
.
Hvitfeldt,
E.
(
2020
)
paletteer: Comprehensive Collection of Color Palettes. R package version 1.2.0.
IPIECA, IOGP, and API
(
2017
)
Guidelines on implementing Spill Impact Mitigation Assessment, SIMA
.
IPIECA/IOGP Oil Spill Response JIP
.
ITOPF
(
2011
)
Use of Dispersants to Treat Oil Spills
,
London
.
Kassambara,
A.
(
2020
)
‘ggpubr': “ggplot2” Based Publication Ready Plots. R package version 0.3.0.
Kleindienst,
S.,
Seidel,
M.,
Ziervogel,
K.,
Grim,
S.,
Loftis,
K.,
Harrison,
S.,
et al. (
2015
)
Chemical dispersants can suppress the activity of natural oil-degrading microorganisms
.
Proc Natl Acad Sci
112
:
14900
14905
.
Klindworth,
A.,
Pruesse,
E.,
Schweer,
T.,
Peplies,
J.,
Quast,
C.,
Horn,
M.,
and
Glöckner,
F.O.
(
2013
)
Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies
.
Nucleic Acids Res
41
:
1
11
.
Lee,
K.,
Nedwed,
T.,
Prince,
R.C.,
and
Palandro,
D.
(
2013
)
Lab tests on the biodegradation of chemically dispersed oil should consider the rapid dilution that occurs at sea
.
Mar Pollut Bull
73
:
314
318
.
Lenth,
R.
(
2020
)
emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.4.7.
Lindstrom,
J.E.
and
Braddock,
J.F.
(
2002
)
Biodegradation of petroleum hydrocarbons at low temperature in the presence of the dispersant Corexit 9500
.
Mar Pollut Bull
44
:
739
747
.
Lozada,
M.,
Marcos,
M.S.,
Commendatore,
M.G.,
Gil,
M.N.,
and
Dionisi,
H.M.
(
2014
)
The Bacterial Community Structure of Hydrocarbon-Polluted Marine Environments as the Basis for the Definition of an Ecological Index of Hydrocarbon Exposure
.
Microbes Environ
29
:
269
276
.
Al Maawali,
W.,
Al Naabi,
A.,
Al Yaruubi,
M.,
Saleem,
A.,
and
Al Maashri,
A.
(
2019
)
Design and Implementation of an Unmanned Surface Vehicle for Oil Spill Handling
.
In
1st International Conference on Unmanned Vehicle Systems-Oman, UVS 2019
.
Mann,
K.H.
and
Lazier,
J.R.N.
(
2013
)
Dynamics of Marine Ecosystems: Biological-Physical Interactions in the Oceans
: Third Edition.
McGenity,
T.J.,
Folwell,
B.D.,
McKew,
B.A.,
and
Sanni,
G.O.
(
2012
)
Marine crude-oil biodegradation: a central role for interspecies interactions
.
Aquat Biosyst
8
:
1
19
.
McKew,
B.A.
and
Smith,
C.J.
(
2015
)
Real-Time PCR Approaches for Analysis of Hydrocarbon-Degrading Bacterial Communities
.
In
Hydrocarbon and Lipid Microbiology Protocols
.
McGenity,
T.J.,
Timmis,
K.N.,
and
Fernandez,
B.N.
(
eds
).
Berlin, Heidelberg
:
Springer Protocols Handbooks
.
NASA
(
2019
)
April 1, 2019 - Phytoplankton Bloom in North Sea.
National Oceanic and Atmospheric Administration
(
2010
)
Characteristics of Response Strategies: A Guide for Spill Response Planning in Marine Environments
.
Nedwed,
T.
and
Coolbaugh,
T.
(
2008
)
Do basins and beakers negatively bias dispersant-effectiveness tests?
In
International Oil Spill Conference - IOSC 2008, Proceedings
.
pp.
835
842
.
North,
E.W.,
Adams,
E.E.,
Thessen,
A.E.,
Schlag,
Z.,
He,
R.,
Socolofsky,
S.A.,
et al. (
2015
)
The influence of droplet size and biodegradation on the transport of subsurface oil droplets during the Deepwater Horizon spill: A model sensitivity study
.
Environ Res Lett
10
:.
Oksanen,
J.,
Blanchet,
F.G.,
Friendly,
M.,
Kindt,
R.,
Legendre,
P.,
McGlinn,
D.,
et al. (
2019
)
vegan: Community Ecology Package. R package version 2.5.6.
254
:.
Pedersen,
T.L.
(
2020
)
ggraph: An Implementation of Grammar of Graphics for Graphs and Networks. R package version 2.0.3.
Pedersen,
T.L.
(
2019
)
patchwork: The Composer of Plots
.
Cran
.
Prince,
R.C.,
Coolbaugh,
T.S.,
and
Parkerton,
T.F.
(
2016
)
Oil dispersants do facilitate biodegradation of spilled oil
.
Proc Natl Acad Sci
113
:
E1421
E1421
.
Prince,
R.C.,
Kelley,
B.A.,
and
Butler,
J.D.
(
2015
)
Three Widely-Available Dispersants Substantially Increase the Biodegradation of otherwise Undispersed Oil
.
J Mar Sci Res Dev
06
:
1
4
.
Prince,
R.C.,
McFarlin,
K.M.,
Butler,
J.D.,
Febbo,
E.J.,
Wang,
F.C.Y.,
and
Nedwed,
T.
(
2013
)
The primary biodegradation of dispersed crude oil in the sea
.
Chemosphere
90
:
521
526
.
Prince,
R.C.,
Nash,
G.W.,
and
Hill,
S.J.
(
2016
)
The biodegradation of crude oil in the deep ocean
.
Mar Pollut Bull
111
:
354
357
.
R Development Core Team, R.
(
2011
)
R: A Language and Environment for Statistical Computing
.
Rahsepar,
S.,
Smit,
M.P.J.,
Murk,
A.J.,
Rijnaarts,
H.H.M.,
and
Langenhoff,
A.A.M.
(
2016
)
Chemical dispersants: Oil biodegradation friend or foe?
Mar Pollut Bull
108
:
113
119
.
Ribicic,
D.,
Netzer,
R.,
Hazen,
T.C.,
Techtmann,
S.M.,
Drabløs,
F.,
and
Brakstad,
O.G.
(
2018
)
Microbial community and metagenome dynamics during biodegradation of dispersed oil reveals potential key-players in cold Norwegian seawater
.
Mar Pollut Bull
129
:
370
378
.
Risdon,
G.C.,
Pollard,
S.J.T.,
Brassington,
K.J.,
McEwan,
J.N.,
Paton,
G.I.,
Semple,
K.T.,
and
Coulon,
F.
(
2008
)
Development of an analytical procedure for weathered hydrocarbon contaminated soils within a UK risk-based framework
.
Anal Chem
80
:
7090
7096
.
Ron,
E.Z.
and
Rosenberg,
E.
(
2014
)
Enhanced bioremediation of oil spills in the sea
.
Curr Opin Biotechnol
27
:
191
194
.
Tett,
P.
and
Walne,
A.
(
1995
)
Observations and simulations of hydrography, nutrients and plankton in the southern north sea
.
Ophelia
42
:
371
416
.
Thomas,
G.E.,
Cameron,
T.C.,
Campo,
P.,
Clark,
D.R.,
Coulon,
F.,
Gregson,
B.H.,
et al. (
2020
)
Bacterial Community Legacy Effects Following the Agia Zoni II Oil-Spill, Greece
.
Front Microbiol
11
:
1
15
.
Tremblay,
J.,
Yergeau,
E.,
Fortin,
N.,
Cobanli,
S.,
Elias,
M.,
King,
T.L.,
et al. (
2017
)
Chemical dispersants enhance the activity of oil- and gas condensate-degrading marine bacteria
.
ISME J
11
:
2793
2808
.
Trudel,
K.,
Belore,
R.C.,
Mullin,
J. V.,
and
Guarino,
A.
(
2010
)
Oil viscosity limitation on dispersibility of crude oil under simulated at-sea conditions in a large wave tank
.
Mar Pollut Bull
60
:
1606
1614
.
Venables,
W.N.
and
Ripley,
B.D.
(
2002
)
Modern Applied Statistics with S
( Fourth Edition).
Wang,
W.,
Wang,
L.,
and
Shao,
Z.
(
2010
)
Diversity and Abundance of Oil-Degrading Bacteria and Alkane Hydroxylase (alkB) Genes in the Subtropical Seawater of Xiamen Island
.
Microb Ecol
60
:
429
439
.
Wang,
Y.,
Lau,
P.C.K.,
and
Button,
D.K.
(
1996
)
A marine oligobacterium harboring genes known to be part of aromatic hydrocarbon degradation pathways of soil pseudomonads
.
Appl Environ Microbiol
62
:
2169
2173
.
Wilke,
C.O.
(
2015
)
Cowplot: streamlined plot theme and plot annotations for ggplot2
.
Wilke,
C.O.
(
2020
)
ggtext: Improved Text Rendering Support for “ggplot2”. R package version 0.1.0.
Yakimov,
M.M.,
Timmis,
K.N.,
and
Golyshin,
P.N.
(
2007
)
Obligate oil-degrading marine bacteria
.
Curr Opin Biotechnol
18
:
257
266
.
Zeinstra,
M.,
Brussaard,
C.,
McGenity,
T.J.,
McKew,
B.A.,
Thomas,
G.E.,
Murk,
T.,
et al. (
2020
)
ExpOS'D: Experimental Oil Spill Data-sharing
.