Numerical study of the effects of chemical dispersant on oil transport from an idealized underwater blowout

As part of the response efforts that followed the Deepwater Horizon oil spill event, chemical dispersants were applied at the ocean surface and at the wellhead to increase biodegradation rates and reduce the environmental impact of the spill. Among other factors, the effectiveness of the application hinges on how it modifies complex fluid mechanical transport processes. The breakup of surface slicks and the vertical dilution of oil in the water column depends on the transport of oil droplets by turbulence in the ocean mixed layer. The trapping of oil deep in the water column results from the complex dynamics of multiphase plumes rising in the stratified ocean. We use a framework based on multiscale coupled computational domains to simulate deep-water blowouts that accurately represents these complex processes, tracking the oil from the wellhead to its transport in the ocean mixed layer. Simulation results show that dispersant application at the wellhead can be effective in trapping the oil in deep-water intrusions. For the conditions simulated here, surface application of the dispersant has important effects on the oil transport: after about 40 h, the mean advection velocity of the oil plume is reduced by a factor of four and the horizontal diffusivity is increased by a factor of 10.

Figure 1

(a) Surface oil concentration from the far-field domain just before surface dispersant application with target area indicated by white dashed line. Dot-dashed line indicates position of cuts for panels (b) and (c), and the source location is indicated by a cross. The white dashed line in panel (b) indicates the interface between near-field and far-field domains. Thin gray dashed lines indicate the “velocity-field LES domain”. (b) and (c) Cross section through the source showing the rising oil plume and the core of the bubble plume (white contours) for simulations without and with dispersant application at the wellhead, respectively. Blue and orange lines indicate positions of mean peel and trap heights, respectively.

Figure 2

Hodograph of mean Lagrangian velocity (Eulerian velocity of ocean current plus Stokes drift velocity) illustrating the Ekman spiral. The Lagrangian velocity converges to the background flow velocity (red vector) at the bottom of the far-field domain. The surface current (orange), the mean Ekman transport velocity in the OML (purple), and the mean current in the OML (green) are also indicated. The dot markers in the hodograph are $5\mathrm{m}$ apart with the first one being at $z=-0.5\mathrm{m}$.

Figure 3

Time evolution of vertically integrated oil concentration that was initially inside the rectangular patch marked by the dashed lines (a)–(f). The top row (a)–(c) is for the case without dispersant application, while the middle row (d)–(f) is with dispersant application. The three columns show results 6, 12, and 18 h after application of surface dispersant. The black dot in each panel represents the location of the center of mass of the oil patch. The bottom panel shows the total vertically integrated oil concentration for the entire plume 18 h after dispersant application in the target area.

Figure 4

Time evolution of statistics of oil patches with (red lines) and without (blue lines) application of surface dispersant: depth of the center of mass (a), horizontal advection speed of the center of mass (b), horizontal advection direction of center of mass (c), horizontal patch size (d), and the total apparent diffusivity (solid lines) and its components in $x$ (dashed lines) and $y$ (dashed-doted lines) directions (e).

Figure 5

Apparent diffusivity is shown against the scale of dispersion (defined here as $\ell =3{\sigma }_r$) for the LES cases with and without dispersant (solid circles). In the same panel are also shown the fits of Lawrence et al. [28] using also data from Refs. [26, 27].