538
coupled hydrodynamic data are provided by the American Seas model, which is
based on the National Research Laboratory Navy Coastal Ocean Model (Martin
2000). Nutrient cycling and subsequent primary productivity, as well as organism
distribution, are all affected by the hydrodynamic inputs. Nitrogen is used as the
currency of exchange between functional groups (Link et al. 2011), though several
other nutrients are also tracked.
The GoM Atlantis model uses a total of 91 functional groups: 61 age-structured
vertebrates and invertebrates, 19 non-age-structured invertebrates, 6 primary producers, 3 types of detritus, and 2 bacteria. All vertebrates and some exploited invertebrates that are age-structured are tracked in Atlantis by body weight in mg N and
number of individuals. Non-exploited invertebrates are only tracked as biomass
pools in mg N. Carrion detritus, labile detritus, and refractory detritus are the three
detrital functional groups. Any functional group can add to detrital and nutrient
pools through mortality, waste excretion, and remnants from feeding behavior. The
detrital groups can also be consumed as a defined part of any functional groups’
diet. A full review of the modeling system and equations is available in Link et al.
(2011). The specific GoM implementation of Atlantis and model fitting is available
in Ainsworth et al. (2015). For this research project, Atlantis simulations were run
for 50 years, from 2010 to 2060.
31.2.2 Environmental Drivers of the Oil Spill
Oil concentration estimates were provided by a probabilistic framework for oil
droplet tracking based on an open-source Lagrangian stochastic model, the
Connectivity Modeling System (CMS) (Paris et al. 2012, 2013; Aman and Paris
2013). The most recent CMS simulations of the DWH blowout are described in an
article in preparation (Perlin et al. 2020). Implementation of these oil data in Atlantis
follows the methods of Ainsworth et al. (2018). Briefly, oil estimates were provided
at 1/12th degree grid points and in four depth layers, averaged across 0–20 m,
20–50 m, 50–200 m, and 200–2000 m. Dose-response calculations were made at the
grid point scale and averaged over all Atlantis polygons, adjusting for the proportion
of grid points per polygon that do not contain oil. Transient biomass dynamics were
allowed to settle in the first 110 days of the simulation before introducing oil forcing
to the model, which then lasts for 167 days.
There is not enough available sediment oil data from the field to use directly in
the model as the oil driver for the sediment layer in simulations. In addition, the
precise distribution, concentration, and extent of DWH oil on the seafloor are
unknown. Romero et al. (2015, 2017) provided sediment core concentrations, which
were then compared to estimated time- and depth-integrated water column concentrations from CMS, to establish a ratio between sediment and water column concentrations (Ainsworth et al. 2018). Using an estimated ratio between sediment oil
concentrations and modeled oil values allows for the generation of an estimated,
continuous sediment concentration map used to drive sediment oil toxicity.
L. N. Dornberger et al.
coupled hydrodynamic data are provided by the American Seas model, which is
based on the National Research Laboratory Navy Coastal Ocean Model (Martin
2000). Nutrient cycling and subsequent primary productivity, as well as organism
distribution, are all affected by the hydrodynamic inputs. Nitrogen is used as the
currency of exchange between functional groups (Link et al. 2011), though several
other nutrients are also tracked.
The GoM Atlantis model uses a total of 91 functional groups: 61 age-structured
vertebrates and invertebrates, 19 non-age-structured invertebrates, 6 primary producers, 3 types of detritus, and 2 bacteria. All vertebrates and some exploited invertebrates that are age-structured are tracked in Atlantis by body weight in mg N and
number of individuals. Non-exploited invertebrates are only tracked as biomass
pools in mg N. Carrion detritus, labile detritus, and refractory detritus are the three
detrital functional groups. Any functional group can add to detrital and nutrient
pools through mortality, waste excretion, and remnants from feeding behavior. The
detrital groups can also be consumed as a defined part of any functional groups’
diet. A full review of the modeling system and equations is available in Link et al.
(2011). The specific GoM implementation of Atlantis and model fitting is available
in Ainsworth et al. (2015). For this research project, Atlantis simulations were run
for 50 years, from 2010 to 2060.
31.2.2 Environmental Drivers of the Oil Spill
Oil concentration estimates were provided by a probabilistic framework for oil
droplet tracking based on an open-source Lagrangian stochastic model, the
Connectivity Modeling System (CMS) (Paris et al. 2012, 2013; Aman and Paris
2013). The most recent CMS simulations of the DWH blowout are described in an
article in preparation (Perlin et al. 2020). Implementation of these oil data in Atlantis
follows the methods of Ainsworth et al. (2018). Briefly, oil estimates were provided
at 1/12th degree grid points and in four depth layers, averaged across 0–20 m,
20–50 m, 50–200 m, and 200–2000 m. Dose-response calculations were made at the
grid point scale and averaged over all Atlantis polygons, adjusting for the proportion
of grid points per polygon that do not contain oil. Transient biomass dynamics were
allowed to settle in the first 110 days of the simulation before introducing oil forcing
to the model, which then lasts for 167 days.
There is not enough available sediment oil data from the field to use directly in
the model as the oil driver for the sediment layer in simulations. In addition, the
precise distribution, concentration, and extent of DWH oil on the seafloor are
unknown. Romero et al. (2015, 2017) provided sediment core concentrations, which
were then compared to estimated time- and depth-integrated water column concentrations from CMS, to establish a ratio between sediment and water column concentrations (Ainsworth et al. 2018). Using an estimated ratio between sediment oil
concentrations and modeled oil values allows for the generation of an estimated,
continuous sediment concentration map used to drive sediment oil toxicity.
L. N. Dornberger et al.