60
A. Dore et al.
10.2 Methodology
The Fine Resolution Atmospheric Multi-pollutant Exchange (FRAME) model is a
Lagrangian atmospheric chemistry transport model which uses annually averaged
meteorology (wind direction frequency rose and annual precipitation map) to calculate annual average deposition of nitrogen and sulphur as well as gas and particulate
concentrations. The model uses spatially distributed emissions of NH 3 , NO x and SO 2
and simulates gas to particle conversion through both dry phase and aqueous phase
chemistry. Dry and wet deposition to the surface are simulated using scavenging
coefficients and a canopy resistance parameterisation respectively. The model can
be run at either a 1 km or a 5 km resolution over the British Isles.
The model was run at a 5 km resolution using historical estimates of emissions
of NH 3 , NO x and SO 2 for the UK for the years 1970, 1990, 2010 and future projections of emissions for the year 2030. During more recent years high resolution
(1 km) emissions data have become available and the model was run at 1 km resolution for the years 2009, 2010, 2011, 2012, 2013 and 2014. The results were used
to calculate long-term trends in NH 3 concentrations in the UK and the area of the
country with exceedance of the critical level for NH 3 . Evaluation of model performance was undertaken by comparison with measurements from the United Kingdom
Eutrophying and Acidifying Pollutants monitoring network (https://uk-air.defra.gov.
uk/networks/network-info?view=ukeap).
10.3 Results
The spatial distribution of NH 3 concentrations modelled at a 1 km resolution for the
UK is illustrated in Fig. 10.1 with the correlation with measurements of annually
averaged NH 3 concentration from 87 sites. NH 3 concentrations are highest in the
intensive agricultural regions of England and Northern Ireland and lowest in the
remote upland regions of northern Scotland. The correlation with measurements
gave a Pearson correlation coefficient of r = 0.71 and a normalised mean bias of
0.17. Although the model was run at a relatively high spatial resolution of 1 × 1 km,
there is still considerable sub-grid variability in the NH 3 concentration in the rural
environment due to the uneven spatial distribution of agricultural land and natural
ecosystems. However Hallsworth et al [3] demonstrated that the model achieved a
better correlation with measurements of NH 3 in semi-natural areas with 1 × 1 km
resolution modelled data than with 5 × 5 km resolution data. Although the model
employs relatively simple dynamic, chemical and dry deposition schemes compared
to more complex Eulerian models, it was able to obtain a good correlation with NH 3
concentrations compared with other models as part of a model inter-comparison
exercise [2]. This good performance could be attributed to the high vertical resolution
in the model of 1 m near the surface.
A. Dore et al.
10.2 Methodology
The Fine Resolution Atmospheric Multi-pollutant Exchange (FRAME) model is a
Lagrangian atmospheric chemistry transport model which uses annually averaged
meteorology (wind direction frequency rose and annual precipitation map) to calculate annual average deposition of nitrogen and sulphur as well as gas and particulate
concentrations. The model uses spatially distributed emissions of NH 3 , NO x and SO 2
and simulates gas to particle conversion through both dry phase and aqueous phase
chemistry. Dry and wet deposition to the surface are simulated using scavenging
coefficients and a canopy resistance parameterisation respectively. The model can
be run at either a 1 km or a 5 km resolution over the British Isles.
The model was run at a 5 km resolution using historical estimates of emissions
of NH 3 , NO x and SO 2 for the UK for the years 1970, 1990, 2010 and future projections of emissions for the year 2030. During more recent years high resolution
(1 km) emissions data have become available and the model was run at 1 km resolution for the years 2009, 2010, 2011, 2012, 2013 and 2014. The results were used
to calculate long-term trends in NH 3 concentrations in the UK and the area of the
country with exceedance of the critical level for NH 3 . Evaluation of model performance was undertaken by comparison with measurements from the United Kingdom
Eutrophying and Acidifying Pollutants monitoring network (https://uk-air.defra.gov.
uk/networks/network-info?view=ukeap).
10.3 Results
The spatial distribution of NH 3 concentrations modelled at a 1 km resolution for the
UK is illustrated in Fig. 10.1 with the correlation with measurements of annually
averaged NH 3 concentration from 87 sites. NH 3 concentrations are highest in the
intensive agricultural regions of England and Northern Ireland and lowest in the
remote upland regions of northern Scotland. The correlation with measurements
gave a Pearson correlation coefficient of r = 0.71 and a normalised mean bias of
0.17. Although the model was run at a relatively high spatial resolution of 1 × 1 km,
there is still considerable sub-grid variability in the NH 3 concentration in the rural
environment due to the uneven spatial distribution of agricultural land and natural
ecosystems. However Hallsworth et al [3] demonstrated that the model achieved a
better correlation with measurements of NH 3 in semi-natural areas with 1 × 1 km
resolution modelled data than with 5 × 5 km resolution data. Although the model
employs relatively simple dynamic, chemical and dry deposition schemes compared
to more complex Eulerian models, it was able to obtain a good correlation with NH 3
concentrations compared with other models as part of a model inter-comparison
exercise [2]. This good performance could be attributed to the high vertical resolution
in the model of 1 m near the surface.