FORECASTING DRIFT
515
processes that they include. For a fuller review of the state of oil spill
modeling, see Reed et al., 1999.
3.1
Oil and weathering processes
Oil spilled into the ocean ranges from unrefined crude oil to heavily
refined products, with a corresponding wide range of chemical composition. In accidents, there can be several types of oil product spilled, for
example crude oil cargo and diesel fuel. The chemical composition of
the particular oil type and how it is spilled into the ocean determine to
a large degree how weathering processes will transform the spill (Daling et al., 2003) and, consequently, what kind of remedial action can be
considered.
The main weathering processes are evaporation, emulsification and
natural dispersion:
Evaporation. For some oil types high in volatile fractions (e.g., many
crude oils), evaporation from the surface slick removes a significant portion of the total mass within a short time, while other types (e.g., heavy
fuel oils) lose relatively little to evaporation. Evaporation algorithms
depend on the boiling point of the oil components, the ambient temperature, wind speed, film thickness and exposure time, although there is
debate as to which parameters are important. The algorithms in use
vary widely in computational expense and interpolation into empirical
databases is a common practical approach.
Emulsification and natural dispersion. The uptake of water into the
oil forms an emulsion, which may differ dramatically in drift behavior
from the pure oil. Natural dispersion is the uptake of oil droplets of
diminishing size into the water until they are no longer part of the oil
slick in any practical sense. These two processes, illustrated in Figure
3, are competitors inasmuch as each reduces the rate of the other. In
particular, the rate of natural dispersion is reduced with emulsion formation and, consequently, the lifetime of the slick is extended. Changes
in the slick lifetime, in turn, affect the choice of response and possibility
for environmental impact. All algorithms currently used for modeling
emulsification and natural dispersion are curve fits to empirical data.
Oil spill models must also track changes in the basic physical properties
of the oil, such as density and viscosity, under the weathering process.
These calculations are typically based on empirical tables. Inaccuracies
can have important ramifications for the drift prediction. For example,
the density of oil is close to that of water, and slight increases can result
in increased downward mixing under the action of waves.
Précédent

- 513/573

Suivant