210
the amount of oil trapped in OMAs, allowing estimates of the consequences of the
presence of mineral fines (Lee et al. 2008). However, during the DwH event, MOS
was observed that consisted of oil incorporated into phytoplankton aggregates
(Passow et al. 2012; Passow 2016). Modeling the formation of such large and heterogeneous MOS, which incorporates oil, but does not require oil to form, calls for
information on all particle concentrations and sizes (mineral particles and biological
particles such as phytoplankton and fecal pellets).
Models of marine snow formation and sinking using aggregation theory have
been previously developed to study the export of carbon from the surface to the deep
ocean (Jackson 1990; Burd and Jackson 2009; Jokulsdottir and Archer 2016).
Typically, three collisional processes are considered: Brownian motion, which is the
dominant collision mechanism for particles less than 1 μm in size; fluid shear; and
differential sedimentation where a faster settling particle collides with a slower settling one. For particles of size between approximately 1 and 20 μm, fluid shear and
differential sedimentation can be equally important, but for large aggregates, differential sedimentation is the dominant collisional process (Burd and Jackson 2009;
Lambert and Variano 2016). The theoretical descriptions of these processes are well
known and have been experimentally verified in a wide range of disciplines
(Elimelech et al. 1995; Friedlander 2000; Pruppacher and Klett 2010).
Aggregation theory has been used to model the aggregation of oil droplets with
mineral particles such as clay and silica and derive collision efficiencies (a measure
of particle stickiness) and aggregate fractal dimensions (Sterling et al. 2004). These
studies indicate relatively high efficiencies of ~0.4–0.6, indicating a high stickiness
of oil. Typical stickiness of marine particles (without oil or TEP) is less than 0.01,
whereas stickiness in the presence of TEP varies between 0.1 and 0.8 (Mari et al.
2017). The fractal dimension of the OMAs was relatively high and varied between
2.6 and 3.0. In comparison, the fractal dimension of phytoplankton aggregates is
lower, usually between 1.3 and 1.9 (Logan and Alldredge 1989; Laurenceau-Cornec
et al. 2015). The fractal dimension is an important parameter because it determines
how the porosity of an aggregate varies with its size, thereby affecting particle collision rates and sinking velocities.
A first attempt at modeling MOSSFA used aggregation theory to predict marine
snow formation and settling rates and then calculated oil-scavenging rates using the
abundance of oil droplets of a given size and the marine snow clearance rates
(Francis et al. 2017). This model was used to estimate oil removal rates from surface
waters under different turbulence conditions and phytoplankton concentrations and
showed that a strong phytoplankton bloom was able to remove all the oil from the
surface waters in 0.5–4 days, depending on the turbulence strength. The model predicted the flux of oil and marine carbon in August/September 2010 as measured via
a sediment trap, reasonably well, and indicated that the trap caught only the tail end
of the bloom, missing the first part of the MOSSFA event.
A significant problem with modeling MOS and MOSSFA events is that MOS are
formed from dynamic populations of a variety of different particle types, and it is
unclear how the properties of an aggregate depend on those of the particles that
comprise it. In particular, it is unclear how to estimate the fractal dimension and
stickiness of the resultant MOS aggregate. Traditional marine coagulation models
A. Quigg et al.
the amount of oil trapped in OMAs, allowing estimates of the consequences of the
presence of mineral fines (Lee et al. 2008). However, during the DwH event, MOS
was observed that consisted of oil incorporated into phytoplankton aggregates
(Passow et al. 2012; Passow 2016). Modeling the formation of such large and heterogeneous MOS, which incorporates oil, but does not require oil to form, calls for
information on all particle concentrations and sizes (mineral particles and biological
particles such as phytoplankton and fecal pellets).
Models of marine snow formation and sinking using aggregation theory have
been previously developed to study the export of carbon from the surface to the deep
ocean (Jackson 1990; Burd and Jackson 2009; Jokulsdottir and Archer 2016).
Typically, three collisional processes are considered: Brownian motion, which is the
dominant collision mechanism for particles less than 1 μm in size; fluid shear; and
differential sedimentation where a faster settling particle collides with a slower settling one. For particles of size between approximately 1 and 20 μm, fluid shear and
differential sedimentation can be equally important, but for large aggregates, differential sedimentation is the dominant collisional process (Burd and Jackson 2009;
Lambert and Variano 2016). The theoretical descriptions of these processes are well
known and have been experimentally verified in a wide range of disciplines
(Elimelech et al. 1995; Friedlander 2000; Pruppacher and Klett 2010).
Aggregation theory has been used to model the aggregation of oil droplets with
mineral particles such as clay and silica and derive collision efficiencies (a measure
of particle stickiness) and aggregate fractal dimensions (Sterling et al. 2004). These
studies indicate relatively high efficiencies of ~0.4–0.6, indicating a high stickiness
of oil. Typical stickiness of marine particles (without oil or TEP) is less than 0.01,
whereas stickiness in the presence of TEP varies between 0.1 and 0.8 (Mari et al.
2017). The fractal dimension of the OMAs was relatively high and varied between
2.6 and 3.0. In comparison, the fractal dimension of phytoplankton aggregates is
lower, usually between 1.3 and 1.9 (Logan and Alldredge 1989; Laurenceau-Cornec
et al. 2015). The fractal dimension is an important parameter because it determines
how the porosity of an aggregate varies with its size, thereby affecting particle collision rates and sinking velocities.
A first attempt at modeling MOSSFA used aggregation theory to predict marine
snow formation and settling rates and then calculated oil-scavenging rates using the
abundance of oil droplets of a given size and the marine snow clearance rates
(Francis et al. 2017). This model was used to estimate oil removal rates from surface
waters under different turbulence conditions and phytoplankton concentrations and
showed that a strong phytoplankton bloom was able to remove all the oil from the
surface waters in 0.5–4 days, depending on the turbulence strength. The model predicted the flux of oil and marine carbon in August/September 2010 as measured via
a sediment trap, reasonably well, and indicated that the trap caught only the tail end
of the bloom, missing the first part of the MOSSFA event.
A significant problem with modeling MOS and MOSSFA events is that MOS are
formed from dynamic populations of a variety of different particle types, and it is
unclear how the properties of an aggregate depend on those of the particles that
comprise it. In particular, it is unclear how to estimate the fractal dimension and
stickiness of the resultant MOS aggregate. Traditional marine coagulation models
A. Quigg et al.