6 A Proof-of-Concept for Linking the Global Meteorological Model …
37
6.2 Implementation
6.2.1 Transport
In the MPAS-AQ prototype, all horizontal advection processes are performed in the
meteorological core, MPAS-A, which uses a centroidal Voronoi (nominally hexagonal) mesh. It uses a C-grid staggering, where scalar prognostic variables such as
potential temperature, density, water vapor mixing ratios, and any species concentrations are defined at the cell centers, while wind speed is defined by values perpendicular to cell faces. In our coupled online meteorology-air quality framework,
species concentrations are updated at cell centers after all CMAQ processes (excluding advection). MPAS-A then calculates species flux divergence by summing fluxes
over all faces of a cell, which gives an updated (i.e. next dynamical time-step) value
of concentration due to advection. We emphasize that this procedure is consistent
for all types of scalars in MPAS-A resulting in a completely consistent transport
treatment in our MPAS-AQ model.
6.2.2 Emissions
Emissions for the MPAS-AQ simulations were processed from several underlying
databases and following a similar approach as described in Xing et al. [5] for HCMAQ simulations. Anthropogenic emissions from nine sectors were obtained from
the HTAP_v2 database [1]. Biomass burning emissions were obtained from the
EDGARv4.2 database. Finally, climatological biogenic volatile organic compounds
(VOCs) and lightning-produced nitrogen oxides (NO x ) emissions were obtained from
the GEIA (Global Emission Inventory Activity) database. Vertical allocation is also
considered based on different emission categories. Temporal allocation (weekly and
diurnal in the case of monthly emission totals, and monthly, weekly, and diurnal in
the case of annual emission totals) followed the approach described in Xing et al. [5].
Chemical speciation was based on North American speciation profiles as applied to
the 2010 emission inventories described in Pouliot et al. [3] and Janssens-Maenhout
et al. [1]. To transfer the emissions from their native 0.1° × 0.1° grid to the MPASAQ unstructured mesh, each of the 3,600 × 1,800 grid cells of the gridded emissions
fields were assigned to the closest MPAS-AQ cell based on the latitude and longitude of the MPAS-AQ cell center. If multiple emission grid cells were assigned to
the same MPAS-AQ cell, their emissions were summed. Note that the biogenic and
lightning emissions were first interpolated from 1.0° × 1.0° to 0.1° × 0.1° and then
assigned to the closest MPAS-AQ cells.
37
6.2 Implementation
6.2.1 Transport
In the MPAS-AQ prototype, all horizontal advection processes are performed in the
meteorological core, MPAS-A, which uses a centroidal Voronoi (nominally hexagonal) mesh. It uses a C-grid staggering, where scalar prognostic variables such as
potential temperature, density, water vapor mixing ratios, and any species concentrations are defined at the cell centers, while wind speed is defined by values perpendicular to cell faces. In our coupled online meteorology-air quality framework,
species concentrations are updated at cell centers after all CMAQ processes (excluding advection). MPAS-A then calculates species flux divergence by summing fluxes
over all faces of a cell, which gives an updated (i.e. next dynamical time-step) value
of concentration due to advection. We emphasize that this procedure is consistent
for all types of scalars in MPAS-A resulting in a completely consistent transport
treatment in our MPAS-AQ model.
6.2.2 Emissions
Emissions for the MPAS-AQ simulations were processed from several underlying
databases and following a similar approach as described in Xing et al. [5] for HCMAQ simulations. Anthropogenic emissions from nine sectors were obtained from
the HTAP_v2 database [1]. Biomass burning emissions were obtained from the
EDGARv4.2 database. Finally, climatological biogenic volatile organic compounds
(VOCs) and lightning-produced nitrogen oxides (NO x ) emissions were obtained from
the GEIA (Global Emission Inventory Activity) database. Vertical allocation is also
considered based on different emission categories. Temporal allocation (weekly and
diurnal in the case of monthly emission totals, and monthly, weekly, and diurnal in
the case of annual emission totals) followed the approach described in Xing et al. [5].
Chemical speciation was based on North American speciation profiles as applied to
the 2010 emission inventories described in Pouliot et al. [3] and Janssens-Maenhout
et al. [1]. To transfer the emissions from their native 0.1° × 0.1° grid to the MPASAQ unstructured mesh, each of the 3,600 × 1,800 grid cells of the gridded emissions
fields were assigned to the closest MPAS-AQ cell based on the latitude and longitude of the MPAS-AQ cell center. If multiple emission grid cells were assigned to
the same MPAS-AQ cell, their emissions were summed. Note that the biogenic and
lightning emissions were first interpolated from 1.0° × 1.0° to 0.1° × 0.1° and then
assigned to the closest MPAS-AQ cells.