176
Macondo wellhead during DWH, are based on prior research (Li et al. 2008 and
Fig. 8 therein; Aman et al. 2011; Aman and Paris 2013). The DSD is binned to set
out a log-normal distribution adapted from Paris et al. (2012) by extending the oil
diameter range from 300 μm to 500 μm, resulting in d 50 of 103 μm and 85 μm for
the untreated and chemically treated live oil, respectively. Addition of the chemical
dispersant (i.e., treated oil) reduces the interfacial tension between oil and water
(Rewick et al. 1984) and shifts the mean droplet size (d 50 ) toward smaller values.
Here, because oil concentrations are estimated by post-processing, multiple scenarios of droplet size distribution (DSD) could be carried out using the same model
output, i.e., the trajectories output of the deep-sea blowout control run representing
the DWH spill that starts on April 20, 2010 (DB_control). Fluid flow includes a
release of 3000 oil droplets at the trap height every 2 hours for 87 days until July 15,
2010 – the day of successful containment of the spill using a capping stack. The oil
droplets are tracked for total of 167 days from the initial blowout date. The release
location is 28.736
°
N, 88.365
°
W at 1222 m, i.e., the estimated height of oil and gas
separation 300 m above the wellhead (Socolofsky et al. 2011). Initial droplet diameters are determined at random by the CMS model in the range from 1 μm to
500 μm. The post-processing algorithm then utilized the change-of-variable technique for the probability density functions to simulate various DSDs for the scenarios with (chemically treated) and without (untreated) SSDI. Each droplet
deployed in the CMS model contains three pseudo-components (fractions) accounting for the differential petroleum hydrocarbons volatility as follows: 10% of light
molecular weight with the density of 800 kg/m
3
, 75% of medium molecular weight
with 840 kg/m
3
, and 15% of heavy molecular weight with 950 kg/m
3
density. The
biodegradation rates for the oil fractions are based on laboratory experiments for the
first-order decay at high pressure and are set to 30 h for the light fraction and 40 h
for the intermediate oil fraction (Schedler et al. 2014; Lindo-Atichati et al. 2014).
Half-life decay rate for the heavy fraction is set to 180 h based on the observational
study of the degradation rates in the GoM (Hazen et al. 2010, Supplementary material Table S7). Additional decay of the oil occurred at the surface due to evaporation,
which half-life was set to 250 h in the control case (see discussion in De Gouw et al.
2011 on evaporation rates for different hydrocarbon types).
Ocean hydrodynamic forcing for the present study uses daily output from the
HYbrid Coordinate Ocean Model (HYCOM; Chassignet et al. 2003, Halliwell
2004) for the GoM region on a 0.04 degree horizontal grid and provides diagnostic
model output at 40 vertical levels spanning from the surface down to 5500 m.
HYCOM model employs data assimilation using the Navy Coupled Ocean Data
Assimilation (NCODA; Cummings 2005), which assimilates available satellite
altimeter observations, satellite and in situ sea surface temperature (SST) observations, as well as available in situ temperature and salinity profiles from moored
buoys, XBTs, and Argo floats. HYCOM output variables used for CMS simulations
included horizontal and vertical velocity components, temperature, and salinity.
The experiments included parameterization of the effects of the surface wind
drift, the importance of which was emphasized by Le Hénaff et al. (2012). Wind
stress components from the 0.5 degree Navy Operational Global Atmospheric
N. Perlin et al.
Macondo wellhead during DWH, are based on prior research (Li et al. 2008 and
Fig. 8 therein; Aman et al. 2011; Aman and Paris 2013). The DSD is binned to set
out a log-normal distribution adapted from Paris et al. (2012) by extending the oil
diameter range from 300 μm to 500 μm, resulting in d 50 of 103 μm and 85 μm for
the untreated and chemically treated live oil, respectively. Addition of the chemical
dispersant (i.e., treated oil) reduces the interfacial tension between oil and water
(Rewick et al. 1984) and shifts the mean droplet size (d 50 ) toward smaller values.
Here, because oil concentrations are estimated by post-processing, multiple scenarios of droplet size distribution (DSD) could be carried out using the same model
output, i.e., the trajectories output of the deep-sea blowout control run representing
the DWH spill that starts on April 20, 2010 (DB_control). Fluid flow includes a
release of 3000 oil droplets at the trap height every 2 hours for 87 days until July 15,
2010 – the day of successful containment of the spill using a capping stack. The oil
droplets are tracked for total of 167 days from the initial blowout date. The release
location is 28.736
°
N, 88.365
°
W at 1222 m, i.e., the estimated height of oil and gas
separation 300 m above the wellhead (Socolofsky et al. 2011). Initial droplet diameters are determined at random by the CMS model in the range from 1 μm to
500 μm. The post-processing algorithm then utilized the change-of-variable technique for the probability density functions to simulate various DSDs for the scenarios with (chemically treated) and without (untreated) SSDI. Each droplet
deployed in the CMS model contains three pseudo-components (fractions) accounting for the differential petroleum hydrocarbons volatility as follows: 10% of light
molecular weight with the density of 800 kg/m
3
, 75% of medium molecular weight
with 840 kg/m
3
, and 15% of heavy molecular weight with 950 kg/m
3
density. The
biodegradation rates for the oil fractions are based on laboratory experiments for the
first-order decay at high pressure and are set to 30 h for the light fraction and 40 h
for the intermediate oil fraction (Schedler et al. 2014; Lindo-Atichati et al. 2014).
Half-life decay rate for the heavy fraction is set to 180 h based on the observational
study of the degradation rates in the GoM (Hazen et al. 2010, Supplementary material Table S7). Additional decay of the oil occurred at the surface due to evaporation,
which half-life was set to 250 h in the control case (see discussion in De Gouw et al.
2011 on evaporation rates for different hydrocarbon types).
Ocean hydrodynamic forcing for the present study uses daily output from the
HYbrid Coordinate Ocean Model (HYCOM; Chassignet et al. 2003, Halliwell
2004) for the GoM region on a 0.04 degree horizontal grid and provides diagnostic
model output at 40 vertical levels spanning from the surface down to 5500 m.
HYCOM model employs data assimilation using the Navy Coupled Ocean Data
Assimilation (NCODA; Cummings 2005), which assimilates available satellite
altimeter observations, satellite and in situ sea surface temperature (SST) observations, as well as available in situ temperature and salinity profiles from moored
buoys, XBTs, and Argo floats. HYCOM output variables used for CMS simulations
included horizontal and vertical velocity components, temperature, and salinity.
The experiments included parameterization of the effects of the surface wind
drift, the importance of which was emphasized by Le Hénaff et al. (2012). Wind
stress components from the 0.5 degree Navy Operational Global Atmospheric
N. Perlin et al.