Air quality expert group - Fine particulate matter (PM2.5) in ... - Defra
Air quality expert group - Fine particulate matter (PM2.5) in ... - Defra
Air quality expert group - Fine particulate matter (PM2.5) in ... - Defra
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>PM2.5</strong> <strong>in</strong> the UK<br />
150<br />
6. The non-l<strong>in</strong>earity of the model response to emission reduction is the result of<br />
complex <strong>in</strong>teractions. It highlights the importance of <strong>in</strong>troduc<strong>in</strong>g the emissions<br />
<strong>in</strong>to the model us<strong>in</strong>g realistic temporal profiles.<br />
A2.2: WRF/CMAQ applications at the Centre for<br />
Atmospheric and Instrumentation Research (CAIR),<br />
University of Hertfordshire<br />
7. CMAQ is not currently be<strong>in</strong>g used as a policy tool <strong>in</strong> the UK but is be<strong>in</strong>g<br />
considered for regulatory applications as part of the Comparison of Simple<br />
and Advanced Regional Models (CREMO) funded by the Environment Agency.<br />
<strong>Defra</strong> is commission<strong>in</strong>g work to assess the potential of CMAQ to meet its policy<br />
needs. The <strong>in</strong>tention of this work will be to develop an operational version of<br />
the model for UK policy applications. This and the follow<strong>in</strong>g section provide<br />
examples of applications for the years 2003, 2005 and 2006, which formed part<br />
of the AQMEII, MEGAPOLI (FP7), <strong>Defra</strong> model <strong>in</strong>ter-comparison and CREMO<br />
projects. Further details on the model set-up and applications can be found <strong>in</strong><br />
Yu et al. (2008) and Chemel et al. (2010).<br />
8. The focus of research has been on <strong>PM2.5</strong> and modelled results are compared<br />
with the observations from Harwell, London North Kens<strong>in</strong>gton and London<br />
Bloomsbury. Overall, CMAQ underestimates <strong>PM2.5</strong> concentration at the<br />
selected sites but reproduces the temporal variations. There are important<br />
considerations when compar<strong>in</strong>g modelled and measured data for the results<br />
presented here. First, the spatial grid resolution is 18 km x 18 km and is not<br />
optimum for simulat<strong>in</strong>g pollutants that are, at least partly, generated on<br />
smaller, local scales. This set-up has been used as a pragmatic approach to<br />
develop a regional configuration to model hourly concentrations over multiple<br />
years. Second, previous results (not shown here) have <strong>in</strong>dicated sensitivity of<br />
<strong>PM2.5</strong> concentrations to boundary conditions, for example, <strong>in</strong> the comparison<br />
conducted with boundary conditions from GEMS and from GEOS-Chem (global<br />
air pollution model). Third, the treatment of aerosols is part of ongo<strong>in</strong>g model<br />
developmental work and significant changes will be <strong>in</strong>troduced <strong>in</strong> the next<br />
version of CMAQ. The results shown <strong>in</strong> Figure A2.2.1 are annual <strong>PM2.5</strong> modelled<br />
results for 2003 us<strong>in</strong>g CMAQ version 4.6 and annual <strong>PM2.5</strong> modelled results for<br />
2006 us<strong>in</strong>g CMAQ version 4.7.1. For 2003 the grid spac<strong>in</strong>g was 15 km. The<br />
2006 run conducted at 18 km grid spac<strong>in</strong>g was part of the AQMEII model <strong>in</strong>tercomparison<br />
for a doma<strong>in</strong> which covered all of Europe.<br />
9. The daily average <strong>PM2.5</strong> concentrations were calculated for a rural background<br />
station Harwell (HAR) for 2003 and 2006, and for urban background stations<br />
London North Kens<strong>in</strong>gton (LNK) for 2006 and London Bloomsbury (LBB)<br />
for 2003. Modelled <strong>PM2.5</strong> ground (first level) results are compared with<br />
measurements from the selected stations for the year 2006 and 2003 for the<br />
appropriate stations <strong>in</strong> Figure A2.2.1. This figure shows that overall CMAQ<br />
values reproduce the temporal variations for all three stations but show about<br />
a 30-40% underestimation especially for the urban stations. Further analysis is<br />
required to understand the model response dur<strong>in</strong>g peak occurrences, where it<br />
correctly captures the tim<strong>in</strong>g but not necessarily the magnitude of the event.<br />
This will <strong>in</strong>clude exam<strong>in</strong>ation of the local and regional nature of <strong>PM2.5</strong> and its<br />
precursors, as well as the govern<strong>in</strong>g meteorological processes (simulated with<br />
WRF).