Assessment of Oil Spill Dispersion and Weathering Processes in Saronic Gulf

Introduction

A study by [1] shows that rising global demand and declining land-based oil reserves have led to a major increase in offshore oil and gas production since the 1990s. At the same time, improvements in oil transport-such as supertankers and large pipeline networkshave helped move oil across oceans more efficiently. However, this growth also increases the risk of oil spills, which can seriously harm marine life and have long-lasting environmental, social, and economic impacts.

Oil spills can happen in many ways: from natural leaks, oil transport, drilling, or accidents involving ships, pipelines, or rigs. Smaller spills are usually easier to manage, but large spills are much harder to control and more damaging. Even though recent studies suggest a drop in major spills (see Table 1 on Mediterranean incidents), the public often focuses on big disasters and overlooks the many smaller spills that happen daily [2]. According to ITOPF and the European Space Agency, more than 80% of oil spill incidents since 1970 have been small (under 7 tons). Still, around 250,000 tons of oil are lost each year during ship operations, with another 120,000 tons spilled near refineries and terminals. Unfortunately, data on accidental spills is often missing or incomplete, highlighting the urgent need for better spill detection and monitoring systems.

Table 1: Recorded incidents in the Mediterranean Sea.

Oil pollution doesn’t just affect the sea surface-it can also reach deeper waters, worsening its environmental impact. As oil drilling moves into deeper areas and pipelines are built at greater depths, the risk of accidents like blowouts and leaks increases [3]. Major deep-water spills include the 2010 Deepwater Horizon disaster in the Gulf of Mexico, which released an estimated 492,000 to 627,000 tons of oil, and the 2011 Penglai 19-3 spill in China’s Bohai Sea, which released about 200 tons [4].

How an oil spill spreads and changes in the ocean depends on many interacting physical, chemical, and biological processes. Key factors include the type of oil, ocean conditions (like wind, waves, and currents), and how the oil is released-suddenly or continuously, at the surface or from deep underwater. Oil undergoes weathering processes such as spreading, evaporation, emulsification, dissolution, photo-oxidation, biodegradation, and settling. These, along with physical forces like mixing, transport, and resurfacing, determine how the spill moves and how long its effects last [5].

When oil is released into the sea, it quickly spreads into a thin layer on the surface [6]. Oil spill models use this spreading behavior to estimate the thickness or surface area of the slick, which is key for predicting how the oil will move and change. The speed and extent of spreading depend on factors like water temperature, and the oil’s viscosity and density [7]. Evaporation happens when the lighter parts of the oil turn into gas and escape into the air-usually within a few hours-reducing the spill’s toxicity [8]. Emulsification occurs when waves mix water into the oil, forming a thick, sticky “mousse.” Thicker oils tend to form more mousse, which slows evaporation [9]. Dissolution allows oil to mix into the water, either from the surface or from small droplets. Although both evaporation and dissolution remove oil, evaporation usually happens faster and affects more of the spill [10]. Biodegradation, once thought to take months, can now begin within a week in some cases, as seen in the Deepwater Horizon spill [11, 12]. Photo-oxidation occurs when sunlight chemically alters the oil, producing water-soluble compounds [12]. Sedimentation happens when oil droplets sink due to increased density or by being eaten and excreted by marine organisms [13].

The Weather Research and Forecasting (WRF) model stands as a leading-edge numerical weather prediction and regional atmospheric model, crafted for both research and operational applications. Developed through a collaborative effort among various agencies and institutions, including National Centre for Atmospheric Research (NCAR), National Oceanic and Atmospheric Administration (NOAA), the Department of Defense, and university researchers, WRF embodies a versatile and efficient codebase adaptable to diverse computing environments. Its modular structure and wide array of physics and dynamics options make it suitable for an extensive range of applications, from real-time weather forecasting to regional climate simulations and air quality modeling. With its widespread adoption globally, WRF continues to serve as a vital tool for advancing our understanding and prediction of mesoscale weather phenomena [14].

Several studies have used the WRF model to better understand wind patterns and improve weather forecasting in coastal and offshore regions. For example, [15] used WRF to study wind behavior over the eastern Mediterranean and compared model outputs with satellite data from Envisat ASAR. This helped them analyze complex wind features like mountain lee waves. [16] showed that combining WRF data with satellite observations improved alignment with buoy measurements, aiding wind energy planning along the Iberian coast. [17] created detailed wind records (SeaWind I and II) using WRF for the Mediterranean, refining the model through sensitivity tests and validating results with buoys and satellite data. [18] tested eight WRF setups across the Northern Sahara and Mediterranean, finding that grid nudging improved wind and temperature predictions [19] used WRF to study wind energy potential off the coast of Southern Brazil, finding good agreement with the Blended Sea Winds product. [20] applied WRF to examine weather changes caused by the 2010 Deepwater Horizon oil spill, underlining the importance of such modeling for cleanup safety and environmental management.

OpenDrift is an open-source, Python-based modeling tool developed by the Norwegian Meteorological Institute for simulating how particles move in air or water [21]. Its flexible design allows it to be used in oil spill tracking, search and rescue, marine biology, and more. [22] used OpenDrift with high-resolution data to study the 2010 Deepwater Horizon oil spill. They found that river outflows had a major impact on where and how much oil reached shore, showing the importance of including river fronts in oil spill models. [23] used OpenDrift to study oil movement over six years, focusing on how ocean currents like the Loop Current affect oil transport in the Straits of Florida. [24] applied OpenDrift with a detailed ocean model in New Zealand’s Bay of Plenty. Using data from 2003– 2004, the model successfully reproduced coastal currents and showed that local winds strongly influenced nearshore movement, especially through wind-driven upwelling in certain areas.

This study is organized as follows: Section 2 outlines the methodology, including a description of the Saronic Gulf study area, an overview of the OpenOil model, integration of real-time weather and ocean data, and details of the simulation setup. Section 3 presents the results, covering three WRF spatial resolution scenarios, the influence of biodegradation, and changes in oil mass and properties. Section 4 provides conclusions, highlights key findings, and offers recommendations for future research and model improvements.

Methodology

The methodology section starts with an introduction to the OpenOil model and its requirements. It then provides an overview of the Saronic Gulf, followed by a discussion of the physical properties, governing equations, and components. Finally, it outlines the simulation setup and experiment parameters.

OpenOil model and its elements

This analysis examines a hypothetical oil spill near the major Greek harbor of Piraeus in the Saronic Gulf. The simulation uses the OpenOil model, which combines oil spill transport and weathering with real-time meteorological and oceanographic data. OpenOil is part of the open Drift framework [21] and models oil as individual particles with specific properties like mass, viscosity, and density, known as Lagrangian elements.

The movement of each oil particle is affected by factors such as currents, wind, and Stokes drift, along with random walk models to simulate diffusion from turbulence. The model also accounts for physical processes like wave entrainment [25], turbulencedriven vertical mixing [26], oil resurfacing due to buoyancy [27], and emulsification [28]. Oil resurfacing is based on oil density and droplet size, with sinking velocity determined by Stokes’ Law, making the model sensitive to the initial oil droplet size distribution [22].

Physical OpenOil parameters

The current simulations incorporate a comprehensive understanding of various physical processes governing oil transport. These include (a) horizontal transport, which accounts for ambient horizontal currents, the impact of wave-induced Stokes drift, and the wind drift. (b) Vertical transport and mixing are also crucial factors, involving phenomena such as waves breaking in the open sea, particles resurfacing due to buoyancy, and their movement under vertical turbulence.

Horizontal transport

Oil elements are subject to advection by currents, wind, and waves, which collectively influence their horizontal motion. Whether submerged or at the surface, oil particles follow the ambient current, experience wind drift – crucial for their horizontal movement – which typically ranges from approximately 1% to 6% of the surface wind speed, often around 3%. Additionally, they are affected by the surface Stokes drift, with its profile calculated by [29] based on the Phillips spectrum.

Vertical transport

A) Wave entrainment: Wave entrainment involves repositioning oil droplets from the surface slick into the water column, influenced by wave energy, oil properties, and environmental conditions [30]. Recent studies have parameterized wave entrainment, considering oil slick characteristics, water density, and energy from breaking waves [31]. The rate of oil entrainment, Q, is described by Equation (1):

where, We (Weber number) and Oh (Ohnesorge number) represent the balance of inertial, surface tension, and viscous forces, and bw F^bw is the fraction of the sea surface with breaking waves, as discussed in Li et al. (2017) and [32].

B) Oil droplet distribution: Wave entrainment affects the size distribution of oil droplets within the water column. Droplets range from 1 μm to 1 mm, with size distributions described by number or volume distributions [33]. The volume distribution, often used in modeling, is represented by the median droplet diameter D₅₀V (Eq. 2):

where, do is the Rayleigh–Taylor instability maximum diameter, re = 1.791, ρ = 0.46, and q = -0.518 are empirical coefficients [32]. The size distribution follows a log-normal pattern and varies with weather and emulsification rates [22].

C) Oil resurfacing: Buoyancy affects the movement of oil droplets, with the terminal velocity dependent on droplet size and density differences between oil and water [34]. The vertical terminal velocity w is given by equation (3):

where, r is the droplet radius, g is the gravity acceleration,

D) Turbulent mixing: Turbulent mixing redistributes oil droplets vertically, influenced by wind speed, current shear, stratification, and wave energy dissipation. The intensity of turbulent mixing is characterized by a vertical eddy diffusivity coefficient K(z) [35]. This coefficient governs turbulent diffusion, described by Equation 4:

Oil weathering processes

OpenOil uses advanced models for oil weathering, including evaporation and emulsification, based on oil properties from the Oil Library (ADIOS) software by NOAA. Evaporation rates vary by oil type and are influenced by factors like wind speed. Some oils can evaporate completely in a few hours, while others may not evaporate at all. Lighter oils tend to evaporate faster, with 20- 40% evaporating in the first 6-12 hours [34]. Evaporation and emulsification affect oil density, viscosity, and oil-water interfacial tension, which in turn influence droplet size distribution [22].

A) Evaporation: Regarding oil evaporation, OpenOil adapts its treatment based on the conditions of the oil slick [37]. Under calm conditions where the oil forms a smooth surface, evaporation is modeled using Mackay’s analytical method [38]. However, under rough weather conditions, a more complex approach is employed (ADIOS2), which utilizes a pseudo-component evaporation model [39]. In this model, crude oils and refined products are represented as a small number of discrete, non-interacting components, each with its own vapor pressure. The evaporation rate for each component is determined by factors such as wind speed, slick thickness, and the molar fraction and volume of the component. The relative molar volume of each component is estimated based on empirical correlations with the boiling point of alkanes, allowing for accurate calculation of vapor pressures using Antoine’s equation [40]

B) Emulsification: While evaporation reduces the volume of the surface slick, emulsification increases it. In a manner similar to evaporation, the emulsification rate according to ADIOS1 is governed by a simple first-order rate law proposed by [41]. This law estimates the rate of change of water content in the oil-water emulsion as a function of wind speed, current water fraction, and the maximum water fraction of the emulsion. Emulsification tends to increase the water content over time, with typical values for the emulsification rate constant ranging from 1 to 2 μs/m2 of the slick surface. As mixing progresses, the droplet size distribution shifts towards smaller droplets while the total water content remains constant [37].

On the other hand, ADIOS2 considers the rate of emulsion formation using a first-order rate law in the interfacial area, rather than water content [42]. This formulation, proposed by Eley et al. [43], describes the rate of change of the interfacial area in terms of an interfacial parameter (kS), which is sensitive to wave energy. The interfacial area is related to the oil-water interfacial area and the maximum interfacial area, with the water fraction being linked to the interfacial area and the average water droplet diameter.

C) Biodegradation: Biodegradation stands as a pivotal natural process capable of mitigating the environmental repercussions of marine oil spills over the long haul. The rate of biodegradation for oil hinges on several factors including the composition of petroleum hydrocarbons, temperature, microbial species present, and the availability of oxygen and nutrients. The biodegradation algorithm integrated into OpenOil draws from the work of Adcroft et al. [44], asserting that the biological breakdown of oil is primarily influenced by temperature. Specifically, the full biodegradation period (R^-1 ), measured in days, is determined by the Equation (5):

where, R^-1 denotes the time required for oil and T the water temperature, in both dissolved and undissolved phases, to be colonized by bacteria and microorganisms, and for the most resilient compounds within the oil to undergo metabolism.

Study Area and simulation setup

The Saronic Gulf is nested in the west-central Aegean Sea, bordered by the Attica peninsula to the north and east, and the Peloponnese to the southwest. It shares its boundaries with the Aegean Sea to the south and southeast, boasting numerous islands and islets, notably Salamina, Aegina and Hydra, depicted in Fig. 1. Spanning approximately 270 km of coastline, the gulf encompasses a sea surface area of around 2890 km2, with an average water depth of roughly 100 meters [45].

To improve accuracy, the OpenOil model uses atmospheric data from the WRF model and ocean data from the CMEMS database. This combination provides a more realistic simulation of oil spill behavior under different environmental conditions. The WRF model (version 4.3) [46] simulates atmospheric conditions with nested domains: a large-scale domain (25 km resolution), an intermediate domain (5 km), an inner domain (1 km), and a Large Eddy Simulation (LES) domain (200 m resolution). It uses various parameterizations, including the two-moment Morrison scheme for microphysics [47], the RRTMG scheme for radiation [48], the Betts-Miller-Janjic (BMJ) scheme for convection [49], the Mellor- Yamada-Janjic (MYJ) scheme for the Planetary Boundary Layer [50], and the Noah Land model. For higher resolutions (1 km and above), convective and PBL schemes are turned off, as the model can resolve these processes directly. The Global Forecast System (GFS) provides initial and boundary conditions. The LES output frequency is set to 30 seconds, and simulations run for 7 days, with the LES domain running for only 2 days due to high computational cost. A 6-hour spin-up run is conducted before each simulation. Oceanographic data consists of hydrodynamic conditions and wave data. Hydrodynamic data, including ocean currents, come from the Copernicus Marine Environment Monitoring Service (CMEMS) Mediterranean Sea Physics dataset [51], while wave data is from the CMEMS Mediterranean Sea Waves dataset [52].

Oil properties in OpenOil are sourced from the ADIOS Oil database, which contains data on nearly 1,000 oil types [37]. In this study, crude oil is released near Piraeus Harbor over five days, from February 1 to February 6, 2024. The simulation setup is shown in Figure 2, which includes: (a) surface current speed, and (b) significant wave height (VHM0), all derived from the CMEMS datasets. These variables-current speed and wave height-are key factors in oil dispersion. Figures 2a and 2b show a consistent pattern, indicating that current conditions remained stable throughout the simulation period.

The oil type used is “BRENT (AD00187),” with a density of 835.1 kg/m³ and viscosity of 7×10⁻⁶ m²/s at 20°C. A total of 20,000 oil particles are released over a 200 m spill radius, with an initial oil mass of 167,020 kg (200 m³). Table 2 summarizes the initial simulation conditions.

Table 2: Initial simulation conditions and input datasets.

Results

This section presents the simulation results. It begins with three WRF simulations using different horizontal resolutions, compared with data from local weather stations. Next, it examines oil weathering with and without biodegradation, showing how the oil’s mass balance and properties change over time.

Spatial WRF scenarios

Three WRF simulations were performed using horizontal resolutions of 5 km, 1 km, and 200 m. The 5 km and 1 km simulations ran for the full 5-day period, while the 200 m simulation ran for only 2 days due to its higher computational demand. To check the accuracy of the WRF outputs, wind data from three local meteorological stations-provided by the National Observatory of Athens [53]-were used. These stations recorded 10-minute average wind speed and main wind direction (e.g., N, NNW). The stations were located on the islands of Aegina and Salamina (within the Saronic Gulf) and Hydra (just outside the Gulf). Table 3 presents the meteorological information.

Table 3: Meteorological station information.

Figure 3 shows a comparison of model outputs from the 5 km and 1 km simulations with the local station data, presented in alphabetical order.

The differences seen in Figure 3 are mainly due to the unique local conditions around each meteorological station, such as terrain, vegetation, or urban areas, which are not fully captured by the model grid-even at its highest resolution of 200 m. These small-scale (microscale) features can significantly affect wind patterns but are hard to represent accurately in a model [54]. In addition, the WRF model uses complex physical parameterizations to simulate atmospheric processes like radiation, convection, and boundary layer dynamics. These parameterizations come with built-in uncertainties, which can also cause differences between the model results and real observations [14].

In Figure 4, the high-resolution 200 m WRF simulation is included for the first 48 hours of the simulation. This early period is critical, as the effectiveness of emergency response actions taken within the first two days can greatly influence the overall success of recovery efforts and help minimize long-term environmental damage from the spill [55]. The results show very good agreement between the WRF simulations and the observational data-both in terms of pattern behavior and actual values-demonstrating the model’s reliability during this crucial initial response window.

Finally, in Figure 5 the first 48 hours of the oil spill transport simulation is presented, showing the wind speed derived from the WRF model. This figure displays the 5km, 1km and 200m wind speed WRF, along with the trajectories of the oil particles.

Figures 5a, 5b, and 5c show oil drift simulations using three different spatial resolutions: 5 km, 1 km, and 200 m. At 5 km resolution, the oil spread is more dispersed, with fewer active particles (blue) and more stranded ones (red), indicating a broader and less focused coastal impact. At 1 km, the active particles are more concentrated, the number of stranded particles is lower, and the drift path is narrower and better defined. At the finest 200 m resolution, the simulation shows the most concentrated drift path, the highest density of active particles, and the fewest stranded particles, reflecting the highest level of accuracy. All three use the same wind drift settings, but the results differ due to spatial resolution. Finer resolutions (1 km and 200 m) better capture local geography and ocean currents, leading to more accurate modeling of particle movement. These models are also more responsive to environmental changes, but they require more computational resources-especially the 200 m resolution. Higher spatial resolution improves the model’s ability to simulate smallscale wind and ocean features, like eddies and boundary currents, which strongly influence oil movement [56, 57]. This leads to more precise forecasts of where oil might travel and where it may reach the shore.

Impact of Biodegradation on Oil Weathering

In this section, we compare two 3D oil spill simulations: one that includes biodegradation and one that excludes it. Although biodegradation typically unfolds over a longer timescale [32], we observe some notable differences within the 5-day simulation period. The results are presented in Figure 6, which tracks the time evolution of key processes, including the total oil mass distribution (such as evaporated, dispersed, and stranded oil), changes in the oil’s density and viscosity, and environmental factors like wind speed and sea surface current speed.

A) Oil Resurfacing: One of the first noticeable effects is that as wind speed increases after about 28 hours from the oil spill release, a significant portion of the submerged oil resurfaces. This resurfacing is a result of increased turbulence and wave entrainment, which alters the buoyancy and rise velocity of the oil droplets, pushing them back to the surface.

B) Oil Stranding: By the 80-hour mark, the amount of oil stranded on the shore starts to differ between the two scenarios. In the simulation with biodegradation (Scenario a), approximately 12% of the oil mass becomes stranded. However, in the no biodegradation scenario (Scenario b), 15% of the oil is stranded by the same time. This suggests that biodegradation plays a role in reducing the overall amount of oil that adheres to the coast. It is likely due to changes in the physical properties of the oil, such as a breakdown in the water column, which prevents it from reaching the shore.

C) Viscosity Changes (Dynamic Emulsion Viscosity): Over the course of the simulation, the viscosity of the oil changes significantly. For the first 10 hours, the viscosity remains relatively stable at a few hundred cPoise. However, after this period, the viscosity begins to rise. By 36 hours, it increases more sharply:

• In Scenario (a), the viscosity reaches approximately 30,000 cPoise.

• In Scenario (b), it is somewhat lower, around 17,500 cPoise.

This increase in viscosity, particularly in Scenario (a), is attributed to emulsification, a process that is accelerated by biodegradation. Emulsified oil is more challenging to remove from the sea surface because it becomes thicker and more resistant to dispersal. This can complicate cleanup efforts, as it may become more difficult to collect or skim the emulsified oil from the water surface [58].

D) Oil Density: Initially, the oil has a density of 835 kg/m³, but this value changes over time. Within the first 5 hours of the spill, the density increases to around 900 kg/m³, and by 48 hours, it continues to rise, reaching approximately 1012 kg/m³. These changes are a result of the oil absorbing water, which makes it denser. As the oil’s density increases, it becomes more likely to sink or remain submerged, depending on its size and weathering state.

E) Environmental Forcing: The influence of the environment also plays a significant role in the oil’s behavior. As the simulation progresses, wind speeds increase, particularly after 28 hours, which contributes to the resurfacing of submerged oil. On the other hand, sea surface currents remain steady at approximately 9 cm/s, driving the horizontal movement of the oil. These forces lead to increased wave entrainment, which causes oil to remain suspended in the water longer. As the oil droplets break into smaller particles, the oil becomes more evenly dispersed, and the vertical movement of the oil is enhanced, depending on the oil’s size and density. These results are closely linked to the wind and ocean currents. Wind and current turbulence modifies the wave entrainment factor (Eq. 1), leading to greater initial dispersion of oil. It also reduces the size of resurfacing oil droplets (Eq. 2), making them remain suspended longer. Additionally, turbulence influences the vertical movement of oil particles (Eq. 3), promoting their rise or fall depending on their size and density, which results in a more dynamic and dispersed distribution of oil in the water column.

The oil mass budget, broken down by physico-chemical processes throughout the simulation, is presented in Table 4. This table shows the proportion of oil mass dispersed, submerged, surface-bound, stranded, and evaporated at each 20-hour interval, with biodegradation included in the first scenario.

Table 4: Oil weathering processes with and without biodegradation.

In the biodegradation scenario, Table 4 shows that 20 hours after the initial oil release, about 50% of the oil mass has evaporated, 2% has biodegraded, and 17% remains on the sea surface, not dispersed into the water column. This suggests that evaporation is a major process in the early stages following the release. As the simulation continues, around 60 hours from the release, the evaporation rate stabilizes at approximately 51%, with 4% of the oil mass biodegraded, and 16% dispersed into the water. By the end of the simulation (120 hours), the evaporation rate remains stable at 51%, the dispersed oil increases to 17%, and the biodegraded portion rises to 7%.

For the scenario without biodegradation, Table 4 shows that after 20 hours, about 32% of the oil mass has evaporated, and 27% remains on the surface. As the simulation progresses (around 60 hours), the evaporation rate stabilizes at 33%, and 24% of the oil particles are dispersed into the water column. By 120 hours, the percentage of dispersed oil particles increases to 25%, with the amount of evaporated oil remaining constant.

The behavior of the particles under different biodegradation conditions reveals distinct patterns over time. In the biodegradation scenario, dispersed oil starts at 11% and increases gradually to 17% over 120 hours. Submerged oil decreases from 18% to 2%, indicating rapid sinking likely due to densification from biodegradation. Surface particles decrease from 17% to 9%, possibly because biodegradation is breaking them down. Stranded particles increase from 2% to 15%, likely due to the accumulation of biodegraded oil on the shores. Evaporated oil remains consistently high at 50-51%.

In the scenario without biodegradation, dispersed oil starts higher at 13% and steadily increases to 25%, suggesting longer persistence in the water column. Submerged oil decreases gradually, indicating slower sinking. Surface particles start at 27% and decrease steadily to 15%, reflecting a slower breakdown without biodegradation. Stranded oil increases from 2% to 22%, showing a steady movement toward the shores. Evaporated oil remains stable at 32-33%.

These differences demonstrate the significant impact of biodegradation on the distribution and fate of the oil particles. Biodegradation accelerates the breakdown and movement of oil, while the absence of biodegradation results in longer persistence of the oil in the water and more accumulation on the shores.

Conclusions

This study examines the release of BRENT crude oil near Piraeus Harbor over a 5-day period, considering two main scenarios: the first involving the use of WRF and OpenOil under three spatial resolution analyses, and the second incorporating biodegradation processes.

1. Outcomes of Different Spatial Resolution Scenarios in Simulating Oil Spill Transport:

• Higher resolution models (1km and 200m) were more sensitive to environmental changes, providing more accurate simulations of oil particle movement and stranding locations. 2. Impact of Biodegradation Process:

• The biodegradation scenario showed higher oil mass evaporation rates, with around 50% of the oil evaporating within 20 hours, compared to only 32% in the scenario without biodegradation.

• Biodegradation significantly influenced the oil’s fate and distribution, with 2% of the oil biodegraded in the first 20 hours, increasing to 7% by the end of the simulation (120 hours).

• In the biodegradation scenario, a larger portion of the oil remained at the sea surface during the initial hours (17%) compared to 27% in the no-biodegradation scenario.

• Over the 5-day period, the proportion of dispersed oil particles increased to 17% with biodegradation, while it reached 25% without biodegradation.

• These results highlight the significant impact of biodegradation on oil fate, influencing evaporation, dispersion, and biodegradation rates over time. Future Work and Development:

• Simulate accidental seafloor oil spills, such as those from the Prinos oil field in the northern Aegean Sea, to evaluate the behavior of oil spills in different environmental conditions.

• Investigate the effects of environmental factors like temperature, salinity, and nutrient availability on oil biodegradation at varying depths in marine environments.

• Explore the integration of more detailed biological processes into the biodegradation model, focusing on microbial communities and metabolic pathways involved in oil degradation.

• Study the combined effects of prolonged biodegradation and other weathering processes like evaporation and dispersion to better understand their joint impact on oil transport and fate.

Acknowledgments

This research was partially developed as part of a deliverable within the framework of the PERIVALLON project.

Author Contributions

V.P. conceived the study and wrote the abstract; V.P and C.A. wrote introduction and results; D.M. prepared WRF atmospheric input data; A.M. and K.V. reviewed the first draft; A.M, K.V and I.G. supervised the methodology and conclusions sections; S.V and I.K. funding acquisition. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

Data Availability Statement

Available upon request.

References

1. ITOPF (2024). Oil Tanker Spill Statistics 2023.
2. Carpenter A, Kostianoy AG (2018) Oil Pollution in the Mediterranean Sea: Part I: The International Context. In D. Barceló & A. G. Kostianoy, The Handbook of Environmental Chemistry 83.
3. Chen H, An W, You Y, Lei F, Zhao Y, Li J (2015) Numerical study of underwater fate of oil spilled from deepwater blowout. Ocean Engineering, 110(Part A): 227-243.
4. Wang Y, Lee K, Liu D, Guo J, Han Q, et al. (2020) Environmental impact and recovery of the Bohai Sea following the 2011 oil spill. Environmental Pollution 263(Part B): 114343.
5. Keramea P, Spanoudaki K, Zodiatis G, Gikas G, Sylaios G (2021) Oil Spill Modeling: A Critical Review on Current Trends, Perspectives, and Challenges. Journal of Marine Science and Engineering 9(2): 181.
6. Simecek Beatty D, Lehr WJ (2017) Extended oil spill spreading with Langmuir circulation. Marine Pollution Bulletin 122(1-2): 226-235.
7. Fingas M (2017) Chapter 10 - In Situ Burning: An Update. In M. Fingas (Ed.), Oil Spill Science and Technology (Second Edition). Gulf Professional Publishing pp. 483-676.
8. Zafirakou A (2019) Oil Spill Dispersion Forecasting Models. Intech Open.
9. Daling PS, Moldestad MØ, Johansen Ø, Lewis A, Rødal J (2003) Norwegian Testing of Emulsion Properties at Sea-The Importance of Oil Type and Release Conditions. Spill Science & Technology Bulletin 8(2): 123-136.
10. Abianeh OS, Chen C (2012) Modelling of evaporation and dissolution of multicomponent oil droplet in shallow water. Advances in Computational Methods in Experimental Heat Transfer 12: 231.
11. McGenity TJ, Folwell BD, McKew BA, Sanni GO (2012) Marine crude-oil biodegradation: a central role for interspecies interactions. Aquatic Biosystems 8(1):10
12. Ward CP, Sharpless CM, Valentine DL, French McCay DP, Aeppli C, et al. (2018). Partial Photochemical Oxidation Was a Dominant Fate of Deepwater Horizon Surface Oil. Environmental Science & Technology 52(4): 1797-1805.
13. Spaulding ML (2017) State of the art review and future directions in oil spill modeling. Marine Pollution Bulletin 115(1-2): 7-19.
14. Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker D, et al. (2021) A Description of the Advanced Research WRF Version 4.3 (No. NCAR/TN-556+STR). University Corporation for Atmospheric Research.
15. Miglietta MM, Zecchetto S, De Biasio F (2010) WRF model and ASAR-retrieved 10 m wind field comparison in a case study over Eastern Mediterranean Sea. Advances in Science & Research 4: 83-88.
16. Salvação N, Bentamy A, Guedes S (2022) Developing a new wind dataset by blending satellite data and WRF model wind predictions. Renewable Energy 198: 283-295.
17. Menendez M, García-Díez M, Fita L, Fernández J, Méndez FJ, et al. (2014) High-resolution Sea wind hindcasts over the Mediterranean area. Climate Dynamics 42(7): 1857-1872.
18. Rizza U, Mancinelli E, Canepa E, Piazzola J, Missamou T, et al. (2020) WRF Sensitivity Analysis in Wind and Temperature Fields Simulation for the Northern Sahara and the Mediterranean Basin. Atmosphere 11(3): 259.
19. Tuchtenhagen P, de Carvalho GG, Martins G, da Silva PE, de Oliveira CP, et al. (2020) WRF model assessment for wind intensity and power density simulation in the southern coast of Brazil. Energy, 190, 116341.
20. Johnson CP (2023) Atmospheric Impacts from the 2010 Deepwater Horizon Oil Spill. LSU Master's Theses 5784.
21. Dagestad KF, Röhrs J, Breivik Ø, Ådlandsvik B (2018) Open Drift v1.0: a generic framework for trajectory modelling. Geosci. Model Dev 11: 1405-1420.
22. Hole LR, Dagestad KF, Röhrs J, Wettre C, Kourafalou VH, et al. (2019). The Deep-Water Horizon Oil Slick: Simulations of River Front Effects and Oil Droplet Size Distribution. Journal of Marine Science and Engineering 7(10): 329.
23. Androulidakis Y, Kourafalou V, Hole LR, Le Hénaff M, Kang H (2020) Pathways of Oil Spills from Potential Cuban Offshore Exploration: Influence of Ocean Circulation. Journal of Marine Science and Engineering 8(7):535.
24. Montaño MM, Suanda SH, Correia de Souza JMA (2023) Modelled coastal circulation and Lagrangian statistics from a large coastal embayment: The case of Bay of Plenty, Aotearoa New Zealand. Estuarine, Coastal and Shelf Science 281: 108212.
25. Li C, Miller J, Wang J, Koley SS, Katz J (2017) Size Distribution and Dispersion of Droplets Generated by Impingement of Breaking Waves on Oil Slicks. Journal of Geophysical Research: Oceans 122(10): 7938-7957.
26. Serra T, Granata T, Colomer J, Stips A, Møhlenberg F, et al. (2003). The role of advection and turbulent mixing in the vertical distribution of phytoplankton. Estuarine, Coastal and Shelf Science 56(1): 53-62.
27. Tkalich P, Chan ES (2002) Vertical mixing of oil droplets by breaking waves. Marine Pollution Bulletin 44(11): 1219-1229.
28. Azevedo A, Oliveira A, Fortunato AB, Zhang J, Baptista AM (2014) A cross-scale numerical modeling system for management support of oil spill accidents. Marine Pollution Bulletin, 80(1-2): 132-147.
29. Breivik Ø, Bidlot JR, Janssen PAEM (2016) A Stokes drift approximation based on the Phillips spectrum. Ocean Modelling 100: 49-56.
30. Lamarre E, Melville W (1991) Air entrainment and dissipation in breaking waves. Nature 351: 469-472.
31. Li Z, Spaulding ML, French-McCay D (2017). An algorithm for modeling entrainment and naturally and chemically dispersed oil droplet size distribution under surface breaking wave conditions. Marine Pollution Bulletin 119(1): 145-152.
32. Keramea P, Kokkos N, Gikas GD, Sylaios G. (2022). Operational Modeling of North Aegean Oil Spills Forced by Real-Time Met-Ocean Forecasts. Journal of Marine Science and Engineering 10(3): 411.
33. Johansen Ø, Reed M, Bodsberg NR (2015) Natural dispersion revisited. Marine Pollution Bulletin 93(1-2): 20-26.
34. Röhrs J, Dagestad KF, Asbjørnsen H, Nordam T, Skancke J, et al. (2018) The effect of vertical mixing on the horizontal drift of oil spills. Ocean Sci 14: 1581-1601.
35. Visser AW (1997) Using random walk models to simulate the vertical distribution of particles in a turbulent water column. Marine Ecology Progress Series 158: 275-281.
36. Thygesen UH, Ådlandsvik B (2007) Simulating vertical turbulent dispersal with finite volumes and binned random walks. Marine Ecology Progress Series 347: 145-153.
37. Lehr W, Jones R, Evans M, Simecek-Beatty D, Overstreet R (2002) Revisions of the ADIOS oil spill model. Environmental Modelling & Software 17(2): 189-197.
38. Stiver W, Mackay D (1984) Evaporation rate of spills of hydrocarbons and petroleum mixtures. Environmental Science & Technology 18(11): 834-840.
39. Jones RK (1997) A Simplified Pseudo-Component Oil Evaporation Model. In Proceedings of the 20th Arctic and Marine Oil Spill Program Technical Seminar, Vancouver, BC, Canada pp. 43-61.
40. Devis Morales A, Rodríguez Rubio E, Rincón Martínez D (2022) Numerical modeling of oil spills in the Gulf of Morrosquillo, Colombian Caribbean. CT&F Ciencia Tecnología Y Futuro 12(1): 69-83.
41. Fingas M (2011) Chapter 8 - Introduction to Spill Modeling. In M. Fingas (Ed.) Oil Spill Science and Technology Gulf Professional Publishing pp. 187-200.
42. Alvarado V, Wang X, Moradi M (2011) Stability Proxies for Water-in-Oil Emulsions and Implications in Aqueous-based Enhanced Oil Recovery. Energies 4(7): 1058-1086.
43. Eley DD, Hey MJ, Symonds JD (1988) Emulsions of water in asphaltene-containing oils 1. Droplet size distribution and emulsification rates. Colloids and Surfaces 32: 87-101.
44. Adcroft A, Hallberg R, Dunne JP, Samuels BL, Galt JA, et al. (2010) Simulations of underwater plumes of dissolved oil in the Gulf of Mexico. Geophysical Research Letters 37(18): L18605.
45. Makatounis PEZ, Stamou AI, Ventikos NP (2023) Modeling the Agia Zoni II tanker oil spill in Saronic Gulf, Greece. Marine Pollution Bulletin, Volume 194, Part B, 115275. ISSN 0025-326X.
46. Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker DM, et al. (2005) A Description of the Advanced Research WRF Version 2 (No. NCAR/TN-468+STR). University Corporation for Atmospheric Research.
47. Morrison H, Curry JA, Khvorostyanov VI (2005) A new double-moment microphysics parameterization for application in cloud and climate models. Part I: Description. Journal of the Atmospheric Sciences 62(5): 1665-1677.
48. Clough SA, Shephard MW, Mlawer EJ, Delamere JS, Iacono MJ, et al. (2005). Atmospheric radiative transfer modeling: A summary of the AER codes. Journal of Quantitative Spectroscopy and Radiative Transfer 91(2): 233-244.
49. Betts AK, Miller MJ (1986) A new convective adjustment scheme. Part II: Single column tests using GATE wave, BOMEX, ATEX and arctic air-mass data sets. Quarterly Journal of the Royal Meteorological Society 112(473): 693-709.
50. Mellor GL, Yamada T (1982) Development of a turbulence closure model for geophysical fluid problems. Reviews of Geophysics 20(4): 851-875.
51. Clementi E, Aydogdu A, Goglio AC, Pistoia J, Escudier R, et al. (2021) Mediterranean Sea Physical Analysis and Forecast (CMEMS MED-Currents, EAS6 system) (Version 1) [Data set]. Copernicus Monitoring Environment Marine Service (CMEMS).
52. Korres G, Oikonomou C, Denaxa D, Sotiropoulou M (2023) Mediterranean Sea Waves Analysis and Forecast (Copernicus Marine Service MED-Waves, MEDWAΜ4 system) (Version 1) [Data set]. Copernicus Marine Service (CMS).
53. Lagouvardos K, Kotroni V, Bezes A, Koletsis I, Kopania T, et al. (2017) The automatic weather stations NOANN network of the National Observatory of Athens: operation and database. Geoscience Data Journal 4(1): 4-16.
54. Sailor DJ (2011) A review of methods for estimating anthropogenic heat and moisture emissions in the urban environment. International Journal of Climatology 31(2): 189-199.
55. Eckroth JR, Madsen MM, Hoell E (2015) Dynamic modeling of oil spill cleanup operations. Proceedings of the 38th AMOP Technical Seminar on Environmental Contamination and Response, Vancouver, BC, Canada pp. 16-35.
56. Warner JC, Armstrong B, He R, Zambon JB (2010) Development of a Coupled Ocean-Atmosphere-Wave-Sediment Transport (COAWST) Modeling System. Ocean Modelling 35(3): 230-244.
57. Huntley HS, Lipphardt BL, Kirwan AD (2011) Surface Drift Predictions of the Deepwater Horizon Spill: The Lagrangian Perspective. In Monitoring and Modeling the Deepwater Horizon Oil Spill: A Record-Breaking Enterprise (eds Y. Liu, A. Macfadyen, Z.-G. Ji and R.H. Weisberg).
58. Fingas M (2013) The Basics of Oil Spill Cleanup (3rd ed.).
59. Google Earth (2024) Saronic Gulf, Pireaus City, Greece.