47 Hierarchical Clustering for Optimizing Air Quality …
303
are removed, with more “signal” of a specific source be present in the hourly data.
This is more visible in species such as SO 2 , primarily emitted from high stacks,
and the least in O 3 , secondarily produced. We have assessed the model capability
to mimic the observation, by taking model results extracted at station locations and
apply the same methodology. We have shown that the model is able to reproduce the
observed clusters particularly well, especially for secondary pollutants such O 3 . We
treated every grid cell of the model output as a potential station location, for different
chemicals, and apply the methodology to generate maps of dissimilarity metric and of
clustering of that metric to provide information for network optimization. These maps
will give areas of representativeness for single stations. Note that this methodology is
not taking into consideration road access or power availability, but the maps generated
can be layered with other information to assist in decision making.
Acknowledgements This project was jointly supported by the Climate Change and Air Quality
Program of Environment and Climate Change Canada, Alberta Environment and Parks, and the
JOSM program.
References
1. B.S. Everitt, S. Landau, M. Leese, D. Stahl. Cluster Analysis, 5th Edition, Wiley Series in
Probability and Statistics, p. 330 (2011)
2. R.A. Johnson, D.W. Wichern. (2007), Applied Multivariate Statistical Analysis. Pearson Prentice
Hall, Pearson Education Inc. Upper Saddle River, NJ, USA
3. M.D. Moran, S. Menard, D. Talbot, P. Huang, P.A. Makar, W. Gong, H. Landry, S. Gravel, S.
Gong, L.-P. Crevier, A. Kallaur, M. Sassi, Particulate-matter forecasting with GEM-MACH15, a
new Canadian air-quality forecast model. In: D.G. Steyn, S.T. Rao (Eds.), Air Pollution Modelling
and its Application XX, Springer, Dordrecht, pp. 2890–292 (2010)
4. E. Solazzo, S. Galmarini, Comparing apples with apples: using spatially distributed time series
of monitoring data for model evaluation. Atmos. Environ. 112, 234–245 (2015)
5. J. Zhang, Q. Zheng, M.D. Moran, P.A. Makar, A. Akingunola, S.-M. Li, G. Marson, M. Gordon,
R. Melick, S. Cho, Emissions preparation for high-resolution air quality modelling over the
Athabasca oil sands region of Alberta, Canada. 21st Intern. Emissions Inventory Conference,
13–17 April, San Diego (2015). http://www.epa.gov/ttn/chief/conference/ei21/session1/zhang_
emissions.pdf
6. I.G. Zurbenko, The Spectral Analysis of Time Series (North-Holland, Amsterdam, 1986), p. 236
303
are removed, with more “signal” of a specific source be present in the hourly data.
This is more visible in species such as SO 2 , primarily emitted from high stacks,
and the least in O 3 , secondarily produced. We have assessed the model capability
to mimic the observation, by taking model results extracted at station locations and
apply the same methodology. We have shown that the model is able to reproduce the
observed clusters particularly well, especially for secondary pollutants such O 3 . We
treated every grid cell of the model output as a potential station location, for different
chemicals, and apply the methodology to generate maps of dissimilarity metric and of
clustering of that metric to provide information for network optimization. These maps
will give areas of representativeness for single stations. Note that this methodology is
not taking into consideration road access or power availability, but the maps generated
can be layered with other information to assist in decision making.
Acknowledgements This project was jointly supported by the Climate Change and Air Quality
Program of Environment and Climate Change Canada, Alberta Environment and Parks, and the
JOSM program.
References
1. B.S. Everitt, S. Landau, M. Leese, D. Stahl. Cluster Analysis, 5th Edition, Wiley Series in
Probability and Statistics, p. 330 (2011)
2. R.A. Johnson, D.W. Wichern. (2007), Applied Multivariate Statistical Analysis. Pearson Prentice
Hall, Pearson Education Inc. Upper Saddle River, NJ, USA
3. M.D. Moran, S. Menard, D. Talbot, P. Huang, P.A. Makar, W. Gong, H. Landry, S. Gravel, S.
Gong, L.-P. Crevier, A. Kallaur, M. Sassi, Particulate-matter forecasting with GEM-MACH15, a
new Canadian air-quality forecast model. In: D.G. Steyn, S.T. Rao (Eds.), Air Pollution Modelling
and its Application XX, Springer, Dordrecht, pp. 2890–292 (2010)
4. E. Solazzo, S. Galmarini, Comparing apples with apples: using spatially distributed time series
of monitoring data for model evaluation. Atmos. Environ. 112, 234–245 (2015)
5. J. Zhang, Q. Zheng, M.D. Moran, P.A. Makar, A. Akingunola, S.-M. Li, G. Marson, M. Gordon,
R. Melick, S. Cho, Emissions preparation for high-resolution air quality modelling over the
Athabasca oil sands region of Alberta, Canada. 21st Intern. Emissions Inventory Conference,
13–17 April, San Diego (2015). http://www.epa.gov/ttn/chief/conference/ei21/session1/zhang_
emissions.pdf
6. I.G. Zurbenko, The Spectral Analysis of Time Series (North-Holland, Amsterdam, 1986), p. 236