Deepwater oil spills: how do plumes spread and negotiate water column?

Later this year, a controlled release of hydrocarbons will be made in the waters off mid-Norway as part of a joint industry project investigating what would happen in the course of a deepwater spill. The experiment will be conducted as part of the DeepSpill JIP (joint industry project), which will provide verification for a deepwater spill model developed at the Sintef research institute in Norway, as well as a separate deepwater blowout model being developed at Clarkson University (joint US government and oil industry funding).

The initiative for DeepSpill originally came from Chevron as part of its two-year involvement in a cooperative research effort with other oil companies and the US Minerals Management Service (MMS), the US government agency that regulates offshore oil and gas development on the US Outer Continental Shelf.

Results unknown

This cooperative research group began investigating the issue in late-1997. The effort was prompted by an awareness that in the event of a well blowout, pipeline rupture or any other cause of a spill in deep water, it was unclear what kinds of response methods would be most effective. Chevron did its own in-house study and set up an expert group which concluded that there were many unknowns regarding the behavior and transport of spills from deep water sources.

"Various computer programs exist for modeling spill behavior, but in deepwater, it would probably be different from shallow water," says Bob Watson, manager of engineering at Norsk Chevron, the DeepSpill project administrator. "What is lacking for these models is verification."

In 1996, also as part of a JIP, Sintef provided verification for a shallow-water spill model which is now widely used by the industry. For various reasons, it also seemed sensible to carry out the deepwater project in Norway, Watson says. Sintef has not only the relevant previous experience, but also a world-class capability in spill modeling. Moreover, the Norwegian authorities take a proactive attitude towards such environmental issues and have proved supportive of the project.

Depth differences

In a shallow-water spill, the gas remains in gaseous form and acts as an engine driving the plume of oil to the surface, more or less vertically above where the spill took place. In the higher pressures and lower temperatures of a deepwater site, it is presumed that the processes that takes place in the wake of a spill will be very different.

The gas will likely form hydrates, in which case its effect in driving the plume upwards will be greatly diminished. The density stratification caused by variations in temperature and salinity, which is not so important in shallow water, becomes significant in deeper waters, says

"We plan to demonstrate plume bending as it meets stratification and begins to travel horizontally," he said. "We expect it to take a minimum of two hours before the oil reaches the surface, which is likely to be a couple of kilometers from the spill site. About 80-90% of the oil is likely to reach the surface."

There are various unknowns to be investigated. "For example, there is the question of the kinetics of hydrate formation, whether it takes seconds or minutes," says Johansen. "The bubbles could also disappear for other reasons than hydrate formation, so we will have to look carefully at the different phases of the gas."

The chosen location is the Helland Hansen ridge in the Norwegian Sea, about 100 km from land, where the water depth is around 1,000 meters, the seabed temperature -1°C, and the pressure about 80 atmospheres.

Three-day trial

Four releases are planned over three days involving volumes of some 120 cu meters of crude oil and 18 cubic meters of liquefied natural gas, which corresponds to 10,000 cu meters of gas at atmospheric pressure and ambient temperature. Two Norwegian crudes will be used, one more prone to emulsification than the other.

A large supply vessel will be used to make the releases and an oceanographic vessel for monitoring. The latter will deploy a current profiler, water sampling equipment and two remotely operated vehicles (ROV).

The primary ROV will be used to track the spill using sonar and video, and the second one for collecting water samples and measuring droplet and bubble sizes.

Air-borne SLAR, UV and IR imaging techniques will be used to survey the slick. Four main categories of data will be collected:

• Metoceanic data for documentation of the experimental conditions
• Hydrocarbons close to the release point, observing the oil droplets, gas bubbles and the transition to hydrate
• Mapping of the plume trajectory and composition
• Surfacing of oil droplets, and the extent and thickness of the surface slick.

A final report is due to be available to participants in draft form by November of this year. The report will include detailed data analysis of the spill results and comparisons between the data and Sintef's DeepBlow transport model.