Search

The Deepwater Horizon oil spill in the Gulf of Mexico in 2010 was the largest in US history, covering more than 1,000 km of shorelines and causing losses that exceeded $50 billion. While oil transformation processes are understood at the laboratory scale, the extent of the Deepwater Horizon spill made it challenging to integrate these processes in the field. This review tracks the Deepwater Horizon oil during its journey from the Mississippi Canyon block 252 (MC252) wellhead, first discussing the formation of the oil and gas plume and the ensuing oil droplet size distribution, then focusing on the behavior of the oil on the water surface with and without waves. It then reports on massive drifter experiments in the Gulf of Mexico and the impact of the Mississippi River on the oil transport. Finally, it concludes by addressing the formation of oil–particle aggregates. Although physical processes lend themselves to numerical modeling, we attempted to elucidate them without using advanced modeling, as our goal is to enhance communication among scientists, engineers, and other entities interested in oil spills.

The freshwater transport (FWT) by the Labrador Current (LC) around the Grand Banks of Newfoundland (GBN) is diagnosed with the 26-year Global Ocean Physical Reanalysis 1/12° data (GLORYS12v1) during 1993 - 2018. The time-mean FWT of the LC above the 1027.25 kg/m3 isopycnal surface is 83.6 mSv (1 mSv = 103m3/s) southward through the Flemish Pass. Among this 83.6 mSv, 42% (35.2 mSv) is exported into the interior of the North Atlantic along the whole pathway of the LC from the Flemish Pass to the Tail of the GBN, with 25.5 mSv by the mean advection and 7.2 mSv by the mesoscale eddy transport. The seasonal and inter-annual variations of the FWT in the east of the GBN are mainly caused by the variation of the horizontal velocity of the LC, and the variation of salinity makes a nontrivial contribution to the variation of the FWT to the north of 45°N. Around the Tail of the GBN, the mesoscale eddies make significant contributions to the time-mean FWT and the seasonal and inter-annual variations of the FWT.

Physical transport processes such as the circulation and mixing of waters largely determine the spatial distribution of materials in the ocean. They also establish the physical environment within which biogeochemical and other processes transform materials, including naturally occurring nutrients and human-made contaminants that may sustain or harm the region’s living resources. Thus, understanding and modeling the transport and distribution of materials provides a crucial substrate for determining the effects of biological, geological, and chemical processes. The wide range of scales in which these physical processes operate includes microscale droplets and bubbles; small-scale turbulence in buoyant plumes and the near-surface “mixed” layer; submesoscale fronts, convergent and divergent flows, and small eddies; larger mesoscale quasi-geostrophic eddies; and the overall large-scale circulation of the Gulf of Mexico and its interaction with the Atlantic Ocean and the Caribbean Sea; along with air-sea interaction on longer timescales. The circulation and mixing processes that operate near the Gulf of Mexico coasts, where most human activities occur, are strongly affected by wind- and river-induced currents and are further modified by the area’s complex topography. Gulf of Mexico physical processes are also characterized by strong linkages between coastal/shelf and deeper offshore waters that determine connectivity to the basin’s interior. This physical connectivity influences the transport of materials among different coastal areas within the Gulf of Mexico and can extend to adjacent basins. Major advances enabled by the Gulf of Mexico Research Initiative in the observation, understanding, and modeling of all of these aspects of the Gulf’s physical environment are summarized in this article, and key priorities for future work are also identified.

Deepwater hydrocarbon releases experience complex chemical and physical processes. To assess simplifications of these processes on model predictions, we present a sensitivity analysis using simulations for the Deepwater Horizon oil spill. We compare the buoyant multiphase plume metrics (trap height, rise time etc), the hydrocarbon mass flowrates at the near-field plume termination and their mass fractions dissolved in the water column and reaching the water surface. The baseline simulation utilizes a 19-component hydrocarbon model, live-fluid state equations, hydrate dynamics, and heat and mass transfer. Other simulations turn-off each of these processes, with the simplest one using inert oil and methane gas. Plume metrics are the least sensitive to the modeled processes and can be matched by adjusting the release buoyancy flux. The mass flowrate metrics are more sensitive. Both liquid- and gas-phase mass transfer should be modeled for accurate tracking of soluble components (e.g. C

We present an analysis of 2225 simulations of artificial oil well blowouts in nearshore and offshore waters of Newfoundland, Canada. In the simulations, we coupled the VDROP-J and TAMOC models to simulate the fate and transport of oil and gas from the release to the sea surface. Simulations were conducted with and without subsea dispersant injection. We analyzed the simulation database to quantify the mass fraction of oil and gas that surfaces, the mass fraction of released benzene that surfaces, and the horizontal offset to the surfacing zone. These data are also synthesized to yield empirical correlations to predict these output metrics from key input parameters. These correlations are summarized in an excel spreadsheet that allows rapid evaluation of spill dynamics with minimal initial knowledge of spill details. We call this tool an offshore response guidance table, which allows exploration of spill dynamics under diverse spill and response options.

Highlights • Pseudo-components for high-pressure deep-water oil spill models • Estimation of Peng-Robinson EOS parameters for distillation-cut pseudo-components • Modeling of the non-ideal chemistry of hydrocarbons in deep waters • New correlations to calculate chemical properties of petroleum fractions • Validated with 614 oils from the ADIOS oil library Abstract Deep-water oil spills represent a major, localized threat to marine ecosystems. Multi-purpose computer models have been developed to predict the fate of spilled oil. These models include databases of pseudo-components from distillation cut analysis for hundreds of oils, and have been used for guiding response action, damage assessment, and contingency planning for marine oil spills. However, these models are unable to simulate the details of deep-water, high-pressure chemistry. We present a new procedure to calculate the chemical properties necessary for such simulations that we validate with 614 oils from the ADIOS oil library. The calculated properties agree within 20.4% with average values obtained from data for measured compounds, for 90% of the chemical properties. This enables equation-of-state calculations of dead oil density, viscosity, and interfacial tension. This procedure enables development of comprehensive oil spill models to predict the behavior of petroleum fluids in the deep sea.

The BP-operated Deepwater Horizon (DWH) oil rig blowout in April 2010 led to the release of about 700,000 tons of crude oil and 250,000 tons of gas into the Gulf of Mexico (GoM) before it was capped. It also led, tragically, to 11 deaths and 16 injuries to people on the rig. This was the first deep, subsurface ocean spill in offshore oil exploration and posed many technical issues as a result. The GoM Research Initiative (GoMRI) was set up with funding provided by BP to improve the science for responding to future spills. Under GoMRI, the Gulf Integrated Spill Research (GISR) consortium was established to increase the understanding of how petroleum fluids in the ocean behave and improve the ability to predict what happens after a spill. GISR's aim was to construct a multiscale nested model suite that covers scales from droplets to an ocean basin, validated by field and laboratory experiments. Thus, we can follow an oil drop from its first release to its arrival on a shoreline, taking into account natural rates of mixing, dissolution, evaporation, and degradation. This was needed, because during the spill, available models all tended to operate at different scales but did not interact. There are also a number of shallow bays along the northern GoM that are of great importance to local fisheries, and these were not included in existing models.

Subsea oil well blowouts and pipeline leaks release oil and gas to the environment through vigorous jets. Predicting the breakup of the released fluids in oil droplets and gas bubbles is critical to predict the fate of petroleum compounds in the marine water column. To predict the gas bubble size in oil well blowouts and pipeline leaks, we observed and quantified the flow behavior and breakup process of gas for a wide range of orifice diameters and flow rates. Flow behavior at the orifice transitions from pulsing flow to continuous discharge as the jet crosses the sonic point. Breakup dynamics transition from laminar to turbulent at a critical value of the Weber number. Very strong pure gas jets and most gas/liquid co-flowing jets exhibit atomization breakup. Bubble sizes in the atomization regime scale with the jet-to-plume transition length scale and follow −3/5 power-law scaling for a mixture Weber number.

Highlights • Biodegradation is a critical process that governs the fate of oil spilled subsea. • Current models use first-order decay of pseudo-components of the oil composition. • We simulated subsea oil spills varying composition, biodegradation rates, and droplet size. • The most important parameter affecting surfacing time and location was droplet size. • The choice of biodegradation model and rates was of secondary importance. Abstract Biodegradation is important for the fate of oil spilled in marine environments, yet parameterization of biodegradation varies across oil spill models, which usually apply constant first-order decay rates to multiple pseudo-components describing an oil. To understand the influence of model parameterization on the fate of subsurface oil droplets, we reviewed existing algorithms and rates and conducted a model sensitivity study. Droplets were simulated from a blowout at 2000 m depth and were either treated with sub-surface dispersant injection (2% dispersant to oil ratio) or untreated. The most important factor affecting oil fate was the size of the droplets, with biodegradation contributing substantially to the fate of droplets ≤0.5 mm. Oil types, which were similar, had limited influence on simulated oil fate. Model results suggest that knowledge of droplet sizes and improved estimation of pseudo-component biodegradation rates and lag times would enhance prediction of the fate and transport of subsurface oil.

This paper reports the results of quantitative imaging using a stereoscopic, high-speed camera system at two natural gas seep sites in the northern Gulf of Mexico during the Gulf Integrated Spill Research G07 cruise in July 2014. The cruise was conducted on the E/V Nautilus using the ROV Hercules for in situ observation of the seeps as surrogates for the behavior of hydrocarbon bubbles in subsea blowouts. The seeps originated between 890 and 1190 m depth in Mississippi Canyon block 118 and Green Canyon block 600. The imaging system provided qualitative assessment of bubble behavior (e.g., breakup and coalescence) and verified the formation of clathrate hydrate skins on all bubbles above 1.3 m altitude. Quantitative image analysis yielded the bubble size distributions, rise velocity, total gas flux, and void fraction, with most measurements conducted from the seafloor to an altitude of 200 m. Bubble size distributions fit well to lognormal distributions, with median bubble sizes between 3 and 4.5 mm. Measurements of rise velocity fluctuated between two ranges: fast-rising bubbles following helical-type trajectories and bubbles rising about 40% slower following a zig-zag pattern. Rise speed was uncorrelated with hydrate formation, and bubbles following both speeds were observed at both sites. Ship-mounted multibeam sonar provided the flare rise heights, which corresponded closely with the boundary of the hydrate stability zone for the measured gas compositions. The evolution of bubble size with height agreed well with mass transfer rates predicted by equations for dirty bubbles.

An integrated observational field effort that makes simultaneous and collocated measurements of turbulence and fine-scale parameters has been conducted near the Deepwater Horizon oil spill site in the northern Gulf of Mexico (GOM). Full water column profiles are collected across the continental slope in July 2013. The observational results suggest that strong turbulence is patchy and mostly measured in the thermocline and deepwater when using the buoyancy Reynolds number, Re b =200 criterion, the boundary between weak and strong turbulence. Bottom enhanced turbulence is often seen on the continental slope. Using the ratio of the turbulent velocity scale and the oil droplets rising velocity, we develop criteria for when turbulence will dominate the movement of oil droplets and when turbulence can be ignored. Based on the data collected, for oil droplets with rising velocity greater than 6×10 −3 m s −1 , the turbulence effect can be ignored on the continental slope of the northern GOM. For oil droplets with rising speed less than 10 −4 m s −1 , their motions will be affected by the turbulent flow at all depths. For oil droplets with rising speed between 10 −4 and 6×10 −3 m s −1 , the role of turbulence will depend on the strength of the local turbulence and water stratification. We also relate turbulent velocity to the size and density of oil droplets by estimating the rising velocity of different size oil droplets due to balance between buoyancy and drag force. Droplet size and density difference are the two critical parameters in determining the role of turbulence.

Understanding the trajectory of oil droplets in crossflow jets is important to estimate the pathways of hydrocarbons and to plan countermeasures. We report experimental results of an oil jet with release velocity around 1.5 m/s in a crossflow of 0.3 m/s. The hydrodynamics of the jet obtained with the Large Eddy Simulation (LES) were used to predict the migration of the oil droplets. Two Lagrangian techniques were explored, one with the inertia of the droplet is considered and the other that treats the droplets as massless particles with rising velocities corresponding to their size. We did not note a large difference between the two approaches. The droplets showed stronger segregation in the vertical direction, which renders the usage of a Gaussian distribution approximation in the vertical inapplicable. The dispersion coefficient at each direction was computed for different-sized droplets. The eddy diffusivity computed based on Boussinesq gradient approximation using the LES data was compared with the dispersion coefficients obtained based on Lagrangian tracking. We also found that droplets 500 μ m and larger escape the vortex while smaller ones get trapped within the vortex. A similar outcome was observed using a vortex trapping function based on inward–outward force balancing at the elevation of the vortex core. The counter-rotating vortex pair (CVP) altered the distribution of droplets of 1 mm and smaller significantly, and bimodal concentration distributions with peaks near the CVP vortex cores and minimum concentration near the center plane were obtained in the lateral–horizontal​ direction. Therefore, measurements of the oil droplet size distribution (DSD) in the center plane of crossflow jets could underestimate the number of small droplets in the whole plume.

In operational oil spill modeling, hydrodynamic models often employ a coarse-resolution grid for computational efficiency. However, this practical grid resolution poorly resolves small-scale flow features, such as starting jet vortices (tidal eddies) that are common at the inlet of bar-built estuaries with narrow inlet channels, particularly where channel dredging and jetties have been employed to aid ship traffic. These eddies influence Lagrangian transport paths and hence the fate of an oil spill potentially entering or leaving an estuary. This research quantifies the effect of tidal eddies on the mixing process and effects at model scales relevant to the operational prediction of oil spills, using the Galveston Bay entrance channel as a study site. Model grid sensitivity was analyzed, yielding an adequate eddy solution at the horizontal grid size of ∼140 m. It is demonstrated that the SUNTANS model at a practical operational grid resolution (∼400 m) captures neither the eddies nor their effects on particle movement, despite showing a satisfactory prediction of net transport through the inlet. The need for subgrid eddy modeling is discussed, and an empirical approach is proposed that can improve oil spill predictions at operational grid resolution scales when results from a high-resolution model are available.

Knowledge of the droplet size distribution (DSD) from the Deepwater Horizon (DWH) blowout is an important step in predicting the fate and transport of the released oil. Due to the absence of measurements of the DSD from the DWH incident, we considered herein hypothetical scenarios of releases that explore the realistic parameter space using a thoroughly calibrated DSD model, VDROP-J, and we attempted to provide bounds on the range of droplet sizes from the DWH blowout within 200 m of the wellhead. The scenarios include conditions without and with the presence of dispersants, different dispersant treatment efficiencies, live oil and dead oil properties, and varying oil flow rate, gas flow rate, and orifice diameter. The results, especially for dispersant-treated oil, are very different from recent modeling studies in the literature.

Development, testing, and application of a deep-sea, high-speed, stereoscopic imaging system are presented. The new system is designed for field-ready deployment, focusing on measurement of the characteristics of natural seep bubbles and droplets with high-speed and high-resolution image capture. The stereo view configuration allows precise evaluation of the physical scale of the moving particles in image pairs. Two laboratory validation experiments (a continuous bubble chain and an airstone bubble plume) were carried out to test the calibration procedure, performance of image processing and bubble matching algorithms, three-dimensional viewing, and estimation of bubble size distribution and volumetric flow rate. The results showed that the stereo view was able to improve the individual bubble size measurement over the single-camera view by up to 90% in the two validation cases, with the single-camera being biased toward overestimation of the flow rate. We also present the first application of this imaging system in a study of natural gas seeps in the Gulf of Mexico. The high-speed images reveal the rigidity of the transparent bubble interface, indicating the presence of clathrate hydrate skins on the natural gas bubbles near the source (lowest measurement 1.3 m above the vent). We estimated the dominant bubble size at the seep site Sleeping Dragon in Mississippi Canyon block 118 to be in the range of 2–4 mm and the volumetric flow rate to be 0.2–0.3 L/min during our measurements from 17 to 21 July 2014.

We compare oil spill model predictions for a prototype subsea blowout with and without subsea injection of chemical dispersants in deep and shallow water, for high and low gas–oil ratio, and in weak to strong crossflows. Model results are compared for initial oil droplet size distribution, the nearfield plume, and the farfield Lagrangian particle tracking stage of hydrocarbon transport. For the conditions tested (a blowout with oil flow rate of 20,000 bbl/d, about 1/3 of the Deepwater Horizon), the models predict the volume median droplet diameter at the source to range from 0.3 to 6 mm without dispersant and 0.01 to 0.8 mm with dispersant. This reduced droplet size owing to reduced interfacial tension results in a one to two order of magnitude increase in the downstream displacement of the initial oil surfacing zone and may lead to a significant fraction of the spilled oil not reaching the sea surface.

As part of the response to the BP Deepwater Horizon spill in 2010, we have been funded by the Gulf of Mexico Research Initiative to construct a nested model suite that can follow an oil particle from its first release to its arrival on a shoreline, taking into account natural rates of mixing and degradation of the oil components. The model suite incorporates (at increasing levels of resolution) a coupled ocean-atmosphere model of the full Gulf of Mexico and North Atlantic Ocean, a deep Gulf of Mexico model, a regional model of the Texas-Louisiana shelf, a 3D, non-hydrostatic bay model, a 3D Navier-Stokes model of the spill plume, and a particle tracking and transformation model for dispersed and dissolved oil and gas fate and transport integrated within the full flow domain. The models are supported by a series of laboratory and field experiments, including studies of single droplets, with and without dispersant, plumes, a deep-sea tracer release experiment and bubble releases to simulate an underwater blowout. The laboratory experiments will improve modeling of small-scale, near-field processes such as bubble and droplet formation, dissolution, droplet-turbulence interaction, and evaporation and dispersion at the air-sea interface. We show how the models are linked and how we are making progress towards the complete nested model suite, which will be available for use in future spills.

Hydrates amass beneath and around production equipment and can form in hydrocarbon–seawater jets/plumes. The sources of hydrocarbons in these hydrates are natural seeps or temporary production system leaks. In this paper, some of (i) the formation parameters for these hydrates and (ii) the impacts on normal production operations and hydrocarbon-spill capture systems are discussed.

A set of weakly dispersive Boussinesq-type equations, derived to include viscosity and vorticity terms in a physically consistent manner, is presented in conservative form. The model includes the approximate effects of bottom-induced turbulence, in a depth-integrated sense, as a second-order correction. Associated with this turbulence, vertical and horizontal rotational effects are captured. While the turbulence and horizontal vorticity models are simplified, a model with known physical limitations has been derived that includes the quadratic bottom friction term commonly added in an ad hoc manner to the inviscid equations. An interesting result of this derivation is that one should take care when adding such ad hoc models; it is clear from this exercise that (1) it is not necessary to do so – the terms can be included through a consistent derivation from the viscous primitive equations – and (2) one cannot properly add the quadratic bottom friction term without also adding a number of additional terms in the integrated governing equations. To solve these equations numerically, a highly accurate and stable model is developed. The numerical method uses a fourth-order MUSCL-TVD scheme to solve the leading order (shallow water) terms. For the dispersive terms, a cell averaged finite volume method is implemented. To verify the derived equations and the numerical model, four cases of verifications are given. First, solitary wave propagation is examined as a basic, yet fundamental, test of the models ability to predict dispersive and nonlinear wave propagation with minimal numerical error. Vertical velocity distributions of spatially uniform flows are compared with existing theory to investigate the effects of the newly included horizontal vorticity terms. Other test cases include comparisons with experiments that generate strong vorticity by the change of bottom bathymetry as well as by tidal jets through inlet structures. Very reasonable agreements are observed for the four cases, and the results provide some information as to the importance of dispersion and horizontal vorticity.

The droplet size distribution of dispersed phase (oil and/or gas) in submerged buoyant jets was addressed in this work using a numerical model, VDROP-J. A brief literature review on jets and plumes allows the development of average equations for the change of jet velocity, dilution, and mixing energy as function of distance from the orifice. The model VDROP-J was then calibrated to jets emanating from orifices ranging in diameter, D, from 0.5 mm to 0.12 m, and in cross-section average jet velocity at the orifice ranging from 1.5 m/s to 27 m/s. The d50/D obtained from the model (where d50 is the volume median diameter of droplets) correlated very well with data, with an R2 = 0.99. Finally, the VDROP-J model was used to predict the droplet size distribution from Deepwater Horizon blowouts. The droplet size distribution from the blowout is of great importance to the fate and transport of the spilled oil in marine environment.

[1] We present laboratory experiments for tidal starting jet vortices forming at idealized barotropic inlets using dye visualization and particle image velocimetry (PIV) of the surface velocity field. Vortices are identified in the PIV data using the local swirl strength, and metrics are calculated for each identified vortex, including position, equivalent circular diameter, maximum vorticity, circulation, and upwelling potential. These quantitative metrics are presented for four different inlet layouts, including narrow and wide barrier islands and short and long jetties. In each case, starting jet eddies initially form attached to the inlet mouth, with a rapid increase in vorticity, circulation, and size as water exiting the inlet flows directly into the starting jet dipole. Once the vortices reach a critical size, they are entrained into the tidal jet and detach from the inlet. As they advect away from the inlet, their size remains steady while the maximum vorticity and circulation gradually decrease because of the effect of bottom friction and the reduction of input to the vortices from the tidal flow as a result of their advection within the tidal jet. Secondary vortices shed from the inlet during the quasi-steady tidal jet are also entrained into the starting jet vortices, decreasing the overall rate at which their vorticity and circulation decay downstream of the inlet. The quantitative results indicate that the starting jet eddies grow, detach from the inlet mouth, and decay at predictable non-dimensional rates and times.

Parameters of the Modified Bartlett-Lewis Rectangular Pulse (MBLRP) stochastic rainfall simulation model were regionalized across the contiguous United States. Three thousand four hundred forty-four National Climate Data Center (NCDC) rain gauges were used to obtain spatial and seasonal patterns of the model parameters. The MBLRP model was calibrated to minimize the discrepancy between the precipitation depth statistics between the observed and MBLRP-generated precipitation time series. These statistics included the mean, variance, probability of zero rainfall and autocorrelation at 1-, 3-, 12- and 24-hour accumulation intervals. The Ordinary Kriging interpolation technique was used to generate maps of the six MBLRP model parameters for each of the 12 months of the year. All parameters had clear to discernible regional tendencies; except for one related to rain cell duration distribution. Parameter seasonality was not obvious and it was more apparent in some locations than in others, depending on the seasonality of the rainfall statistics. Cross-validation was used to assess the validity of the parameter maps. The results indicate that the suggested maps reproduce well the observed rainfall statistics for different accumulation intervals, except for the lag-1 autocorrelation coefficient. The boundaries of the expected residual, with 95% confidence, between the observed rainfall statistics and the simulated rainfall statistics based on the map parameters were approximately ±0.064 mm hr-1, ±1.63 mm2 hr-2, ±0.16, and ±0.030 for the mean, variance, lag-1 autocorrelation and probability of zero rainfall at hourly accumulation levels, respectively. The estimated parameter values were also used to estimate the storm and rain cell characteristics.