Page 2 of 2
In the 1980s, scientists started treating oil slicks as a collection of droplets, where the ability of the oil to spread or float is partially dependent upon droplet size and distribution. Theorists anticipated that high-pressure, low-temperature spill environments would require code modules that can deal with gases in a non-idealized state. As oil-drilling operations moved into waters deeper than 300 meters over the past fifteen years, particularly in the North Sea, off the shores of West Africa and Brazil, and in the Gulf of Mexico, code developers produced new models that incorporate the previous advances but which address deeper environments. Jets from deep blowouts are also treated in the codes as an assortment of droplets, with diameters ranging from micrometers to millimeters in size, but which must expand as they approach shallow water, complicating calculations of the bubbles’ individual drag forces significantly. If, somewhere below the waterline, the drag forces come into balance with a bubble’s buoyancy, then it will remain neutral, going neither up or down until a current or other force pushes it.
The existence of underwater plumes originating from the Deepwater Horizon site agrees well with the blowout model framework. Some of the released materials situate themselves below the surface because as the oil-gas mixture rises through a gradient of pressures it can start to carry, or “entrain,” heavy water up from the depths, and this reduces the buoyancy of the plume. Whereas it might be intuitive that a high fraction of natural gas mixed in with the oil would make the mixture more buoyant, this gas has time to dissolve into the water. Most significantly, hydrates weigh down a plume as it strives for the surface. As Spaulding puts it, “The engine that would drive the blowout to the surface is taken away. ” As the plume reaches shallower depths, the hydrate will decompose into water and gas again. The heavy water gets harder to lift and when it falls out of the jet, it produces secondary oil and gas plumes called “intrusion layers”. If the remaining oil and gas separate in the presence of a cross current that significantly bends the jet, water and fine oil droplets will flow downstream while gas and large oil droplets surge toward the surface. In the Gulf, a tangle of eddies branch off of the Loop Current and interact with deeper currents, potentially providing the cross flow for subsurface plumes.
This is where modeling results become tenuous and, unfortunately, the most critical. There is very little previous experimental data available to verify what happens once the oil and natural gas have begun to follow different courses. It is clear that subsurface intrusion layers will form, but it’s less certain at what heights these layers will materialize and what their concentrations will be. Call them what you will, but one such intrusion has been identified as the 35-kilometer-long, 9-kilometer-wide oil cloud approaching the DeSoto Canyon in the northeastern slope near Alabama and Florida. Water naturally gets denser with depth, but the level of salinity and mixing with currents introduces irregular density variations. No water currents are constant, and while wind patterns dominate superficial currents, deeper flows respond to the topology of the seafloor and get stronger towards the bottom. All these uncertainties are magnified by unclear assessments of just how much oil is flowing out of the broken riser pipe at the bottom of the Gulf of Mexico.
It may seem simple to get a good estimate of the amount of oil escaping into the sea by watching video footage of the gushing well and considering what the apparent rate of escape is and how long it’s been happening. But the reality is more complicated. “There are several factors impacting why flow rate estimates from the video have large uncertainty. The main factor is that no one can tell you the gas-to-oil ratio. The other is that you can only see the outside of the jet—you do not know exactly what the velocity is inside the plume.” From this description offered by Socolofsky, we might visualize the jet as a cylindrical stream of bubbles slipping past one another, each of different sizes and unknown oil, gas, and water concentrations, all buffeted by shifting turbulence as they rise from the seafloor.
“The challenge for three-dimensional modeling is the required fine resolution in the core of the plume. As can be seen in the blowout video, the plume starts with a 20-centimeter diameter, and there is a lot of variability within that core,” Socolofsky says. Unfortunately, once that variability spreads over a water column like that above the Deepwater Horizon leak, modeling it demands more computing power than is readily available. Researchers work around this problem by approximating the plume in only one dimension, then coupling it with a coarse three-dimensional model for the far-field transport of the plume’s oil. The technique is imperfect, but currently the best science has to offer. That may change for any future deep-sea oil blowouts, however. Socolofsky speculates that the current sea-floor disaster could provide better measurements of vertical motions and concentrations of oil, gas, and dispersants, creating a “great advance in plume modeling and reliability of model results.”
But we needn’t rely on disastrous events to progress our understanding of hydrodynamics; some relevant properties scale nicely between micro and macro situations. “A classical example”, says Socolofsky, “is the universality of the entrainment coefficient for a plume, which is the ratio of the speed of fluid entering the plume on the side to the speed of fluid in the center of the plume. This parameter seems to be constant from small laboratory experiments on the order of a few tens of centimeters, all the way up to the scale of volcanic eruptions with scales of several kilometers.” However, scientists still can’t enable the accurate reproduction in a lab setting of region-specific variations in ocean water densities and local currents. For this, controlled field studies are appropriate. Add to the list of things that should have been done prior to the accident: scientific field tests of blowout codes in the Gulf of Mexico for contingency planning.
The deep holds some of our most profound mysteries. In like manner, a disaster originating from the seafloor summons within us a profound sense of uncertainty. Without accurate initial conditions, the rest of the coding exercise can determine little since the buoyancy of the plume and thus its ability to surface depends partly on the shape of the exit hole and the speed with which the mixture leaves it. In other words, things like an irregular cut in the riser pipe made by a broken diamond saw and moderately successful siphoning techniques can have great bearing upon the plume’s fate. Under the correct assumptions, the behavior of a plume should be the same in most underwater situations, but experience has shown that features of the local environment will weigh in on plume behavior—just as the deposited crude becomes inextricable from the regional ecosystem. At present, computer simulations cannot tell our future, but we can do as they do: Let historical currents and real-time information refine our grasp on what is to come.
Originally published June 15, 2010
Page 2 of 2