30 Modelling the Temporal and Spatial Allocation of Emission Data
195
30.3 Emission Disaggregation
30.3.1 Horizontal and Vertical Disaggregation
Emission inventories that are given in gridded format like EDGAR [15] or the
ECLIPSE emissions from the International Institute for Applied System Analysis
(IIASA) [10] are provided on spatial resolutions between approximately 0.1 × 0.1
degrees and 0.5 × 0.5 degrees. Atmospheric chemistry transport models, on the other
hand, often use other spatial resolutions and other map projections. Therefore, the
emission data needs to be re-gridded.
This can be done with rather simple interpolation methods where no other information about the emission sources is considered. Other methods use so called proxydata that contains spatial information related to the emissions for distributing them
on a new grid. Taking for example the residential heating sector, the assumption
here is that the emissions depend on the population density. The emissions from this
sector would first be aggregated for a larger area, e.g. based on annual sales statistics
for energy carriers in a specific country. Then, they would be distributed following
population density data that is available on a very high resolution grid (down to 1 ×
1 km
2 ). Afterwards, the data would be newly aggregated on the grid used for the
chemistry transport model runs, if this is necessary.
Frequently, emissions depend on meteorological quantities, and here in the first
place on ambient temperature. This is especially the case for residential heating
emissions, but it also applies to emissions from agriculture or evaporative emissions
from cars. Consequently, these emissions can be spatially—and temporally—disaggregated following the meteorological data that is necessary to run the chemistry
transport model system. Aulinger et al. [1] presented a study where they disaggregated national emission totals for benzo(a)pyrene according to heating demand which
is obviously temperature dependent.
Emission strengths do not vary only horizontally, there are also large differences
in emission height. Large power plants, industrial facilities, and, e.g., ships have high
stacks that release their emissions in high altitudes. In addition, the exit velocity and
the temperature of the exhaust gas need to be considered when the emissions shall be
vertically allocated. Bieser et al. [3] calculated more than 40,000 vertical emission
profiles for six pollutants and various meteorological conditions using algorithms
implemented in the SMOKE model [7]. These profiles were clustered into 73 groups
using hierarchical cluster analysis. They contain effective emission heights for several
SNAP sectors and can be used as representative emission heights in regional model
applications for Europe. However, this approach could still be improved when more
information about stack properties in a number of countries would be available.
195
30.3 Emission Disaggregation
30.3.1 Horizontal and Vertical Disaggregation
Emission inventories that are given in gridded format like EDGAR [15] or the
ECLIPSE emissions from the International Institute for Applied System Analysis
(IIASA) [10] are provided on spatial resolutions between approximately 0.1 × 0.1
degrees and 0.5 × 0.5 degrees. Atmospheric chemistry transport models, on the other
hand, often use other spatial resolutions and other map projections. Therefore, the
emission data needs to be re-gridded.
This can be done with rather simple interpolation methods where no other information about the emission sources is considered. Other methods use so called proxydata that contains spatial information related to the emissions for distributing them
on a new grid. Taking for example the residential heating sector, the assumption
here is that the emissions depend on the population density. The emissions from this
sector would first be aggregated for a larger area, e.g. based on annual sales statistics
for energy carriers in a specific country. Then, they would be distributed following
population density data that is available on a very high resolution grid (down to 1 ×
1 km
2 ). Afterwards, the data would be newly aggregated on the grid used for the
chemistry transport model runs, if this is necessary.
Frequently, emissions depend on meteorological quantities, and here in the first
place on ambient temperature. This is especially the case for residential heating
emissions, but it also applies to emissions from agriculture or evaporative emissions
from cars. Consequently, these emissions can be spatially—and temporally—disaggregated following the meteorological data that is necessary to run the chemistry
transport model system. Aulinger et al. [1] presented a study where they disaggregated national emission totals for benzo(a)pyrene according to heating demand which
is obviously temperature dependent.
Emission strengths do not vary only horizontally, there are also large differences
in emission height. Large power plants, industrial facilities, and, e.g., ships have high
stacks that release their emissions in high altitudes. In addition, the exit velocity and
the temperature of the exhaust gas need to be considered when the emissions shall be
vertically allocated. Bieser et al. [3] calculated more than 40,000 vertical emission
profiles for six pollutants and various meteorological conditions using algorithms
implemented in the SMOKE model [7]. These profiles were clustered into 73 groups
using hierarchical cluster analysis. They contain effective emission heights for several
SNAP sectors and can be used as representative emission heights in regional model
applications for Europe. However, this approach could still be improved when more
information about stack properties in a number of countries would be available.