Model Observed (a) (b) Fig. 2. 8-h daily maximum ozone concentrations, averaged over all summer days from June 4 to July 31, 2001 Fig. 3. Observed and CMAQ-predicted 8-h daily maximum ozone concentrations from June 1 to July 31, 2001 averaged over Southern Ontario (a) Fig. 5. Simulated surface level pressure pattern and precipitation on June 27, 6:00 UTC (a), July 26, 15:00 UTC (b), and July 12, 9:00UTC (c). The arrow in panel (b) indicates the track of the anticyclone Fig. 1. Location of relevant ozone monitoring sites of the National Air Pollution Surveillance (NAPS) network in Southern Ontario. (b) Fig. 6. CMAQ-predicted 8-h daily maximum ozone concentration averaged over Southern Ontario from June 1 to July 31, 2001. The definition of the scenarios is explained elsewhere. (c) Fig. 4. Daily averaged fraction of cells which the component ‘v’ of the ground level wind direction is > 0 (i.e. southerly wind). Circles are for cells belonging to U.S. territories. Down triangles are for cell belonging to the states of Michigan, Ohio, Pennsylvania or New York. CMAQ-predicted 8-h daily maximum ozone concentrations from June 1 to July 31, 2001 averaged over Southern Ontario (square points) is also shown. LONG RANGE TRANSPORT OF OZONE IN NORTH EASTERN NORTH AMERICA Oscar Gálvez, Fuquan Yang and James J. Sloan Waterloo Center for Atmospheric Sciences, University of Waterloo, Ontario • Abstract • This study focuses on long range transport of ozone as it affects Southern Ontario. The work was carried out using the MM5/SMOKE/CMAQ regional air quality modelling system, together with observational data from monitoring stations located throughout the modelling domain. A back-trajectory cluster methodology was used to evaluate the magnitude of the effects studied. The transport processes depend strongly on meteorological conditions. An analysis of wind direction and cloud cover shows a strong correlation between these parameters and ozone concentration and synoptic pressure patterns were analyzed to examine other meteorological aspects. The fraction of the contribution from ‘background’ ozone was compared with that from long range transport within the region. • Introduction • The sources of Tropospheric ozone can be classified broadly as (Yap et al., 1979): • Natural or ‘background’ tropospheric O3 • Local anthropogenic emissions • Long-range transport and accumulation • Vehicle-related urban plumes • Stratosphere injections • The importance of long range transport has been noted by numerous authors (e.g., Brankov et al., 2003; Brook et al, 2002; Yap et al., 1988), but its magnitude depends strongly on the relative locations of the source and receptor regions and the meteorology connecting them.Since this region contains a population of approximately 11 million, it is important to assess all sources of ozone – particularly long-range transport ‑ in order to design effective emission control strategies. • We present a study on the long range transport of ozone to Southern Ontario. The objectives are to evaluate the magnitude of this process and to assess the ability of the MODEL-3 Community Multiscale Air Quality (CMAQ) system to simulate it. • Target Area • The focus of the study is Southern Ontario (see Figure 1), which is comprised of the Greater Toronto Area (GTA), Central Ontario, Eastern Ontario, and Southwestern Ontario. Over 90 per cent of the Ontario population is included in this area. Contribution from selected emissionsand ‘background’ In order to evaluate the contribution of the anthropogenic emissions from the nearby U.S. states and the ´background´ to the ground level ozone concentration in Southern Ontario, we performed the model over several scenarios, which differ from the ‘base case’ in the following ways: 1) ‘Zero’: No anthropogenic or biogenic emissions. 2) ‘Bio’: No anthropogenic emissions but normal biogenic emissions. 3)‘US’: No anthropogenic emissions over the Canadian part of the model domain. 4) ‘US-Oh’: Scenario 3 plus no anthropogenic emissions from Ohio. 5) ‘US-Ring’: Scenario 4 plus no anthropogenic emissions over Michigan, Pennsylvania and New York. The ozone concentration for Southern Ontario for these scenarios is illustrated in Figure 6. Scenarios 1 and 2 are selected to estimate the ´background´ozone concentration. Small differences appear in high ozone episodes, even thought, the ozone concentrations from scenarios ‘Zero’ and ‘Bio’ are almost identical, indicating that the contribution from the biogenic emissions is not significant for the ‘background’ ozone concentration calculated by CMAQ. The 8-h daily maximum ozone concentration averaged over entire period for the scenario ‘Zero’ and ‘Bio’ is 31 ppb for both cases, while it is 53 ppb for the ‘base case’ which corresponds to the finding of Yap et al. (1988). Back-trajectory analysis Three meteorological dataset were used for this analysis: FNL archive data (190 Km), EDAS data (80 Km), and MM5 data (36 Km). For every site, each of the 58 trajectories for June and July are assigned to a specific transportation sector, if more than a specific percentage of the transport history over 3 days was in only one sector. Trajectories not meeting this criterion were considered unclassifiable.Table 1 shows the average ozone concentration for each sector for different meteorological dataset. The ozone concentration for the East, West and South was considerable higher than for the North sector. Concentration for the North sector is 20-35 ppb (depending on the transportation sector) lower than others, which indicate the relevancy of the transboundary transport. Trajectories coming from the South region recorded highest ozone concentrations: 64-74 ppb (67-79 ppb from CMAQ). East and West sectors produce similar ozone concentrations: 58-64 ppb for the East (68 ppb from CMAQ) and 61-62 ppb for the West (62-63 from CMAQ). Because there is a smaller percentage of trajectories coming from the South and East sectors more discrepancies are produced between observational and CMAQ data, even among different meteorological dataset. The major divergences among scenarios appear when the ozone concentration is high (smog episodes); otherwise the differences are small with the tropospheric ‘background’ being the principal contributor to the predicted ozone. Figure 7a shows that the 8-h maximum average ozone concentration calculated for scenario 3 is more than 90 % of ozone concentration calculated for the ‘base case’, which indicates only a small reduction by neglecting the Canadian anthropogenic emissions. Also, when the anthropogenic emissions for Ohio are eliminated (scenario 4), the ozone decrease is approximately 10 % inside smog episodes (see Figure 7b) compared to scenario 3. The most important reductions in the ozone concentration appear in scenario 5, which predicts reductions of 25 % averaged over entire period or 35 % inside smog episodes. After separating the contribution of the ozone ‘background’, turning-off the Canadian anthropogenic emissions and the closest U.S. states implies a reduction of more than 60 % in the ‘anthropogenic’ ozone. This reduction is approximately 15 ppb in the 8-h maximum average ozone concentration over the entire study period and as high as 27 ppb in the days with 8-h maximum ozone concentration is greater than 65 ppb. Table 1. CMAQ-predicted and observed 8-h daily maximum ozone concentration (ppb), averaged over six sites (described previously) by transportation sectors for different meteorological dataset. Synoptic pattern description The three predominant synoptic pattern which were observed during the studied period are showed in Figure 5: Panel a The amplitude ridge extending from the high pressure system (Bermuda High) over the southeastern quadrant of our modeling domain, which was associated with elevated ozone concentrations in Southern Ontario. This kind of weather pattern situation can be found in the 2 five-day smog episodes starting on June 26 and July 20. It is one of the most frequent occurrences which were associated with high ozone concentration in Southern Ontario (Hogrefe et al., 2004), which produces warm air from the South comes into this region. Consequently, back-trajectories analyses also reveal air parcels originating from the U.S territories during this period and the long range transport of ozone becomes a critical factor for the air quality. Panel b An anticyclone was moving from northwestern Canadian territories to the Great Lakes region, producing northerly flows over Southern Ontario. This pattern was early typified by Heidorn and Yap (1986) as a typical situation in the summers in Southern Ontario. A similar pattern can be found during our study period on June 5 to 9, and also on July 25 to 27, both were associated with very low ozone concentration level. In the days after these periods, the Southern Ontario region was gradually under the influence of the rear side of the anticyclone, also simultaneously, an increase of ozone concentration was recorded. Panel c A low level pressure over the Northeast of the Great Lakes remained nearly stationary. Such weather pattern corresponds to the situation on July 9 to July 15. Back-trajectory analysis showed air masses coming from the North territories, and consequently lower ozone concentrations are recorded in this period of time. Fig 7. Percentage of 8-h daily maximum ozone concentration averaged over Southern Ontario for selected scenarios from ‘base case’ (a) all days from June 4 to July 31, 2001 and (b) only for days with 8-h daily maximum ozone concentration is more than 65 ppb. • Summary • We tested the ability of MM5/SMOKE/CMAQ to simulate the ozone concentration on Southern Ontario, indicating that the model slightly underestimates the mean 8-hour daily maximum ozone concentration over this region, but nevertheless it reproduces with sufficient precision the spatial and temporal distribution of ozone’s patterns. The analysis of the wind field mainly shows the importance of the direction of the wind in the nearby U.S. states for the air quality of Southern Ontario, which suggests a big influence by the emissions from these states. The results of clusters of back-trajectories confirm this hypothesis, showing an increase of approximately 30 ppb in the maximum ozone concentration when the air masses come from U.S. The analysis of the sea-surface level pressure reveals that high ozone concentrations usually correspond with the amplitude ridge extending from the Bermuda High, which is associated with southerly winds. • In order to estimate the contribution of ‘background’, local and long-range transport ozone to the air quality of Southern Ontario, CMAQ model was performed in several scenarios. The results reveal that ‘biogenic’ emissions do not make significant contributions to the ‘background’ ozone concentration, which was estimated to be approximately 31 ppb for Southern Ontario, and it agrees with the results of other authors. Local emissions are not significant for Southern Ontario, but reductions of more than 60 % are observed if selected U.S. anthropogenic emissions are eliminated. • Acknowledgements • Financial support for this work was provided by Secretaría de Estado de Educación y Universidades de España and Fondo Social Europeo, the Natural Sciences and Engineering Research Council of Canada, Ontario Power Generation, and Ontario Research and Development Challenge Fund. We wish to acknowledge Mr. Jonatan Aronsson for technical assistance. • References • Brankov, E., Henry, R.F., Civerolo K.L., Hao W., Rao, S.T., Misra P.K., Bloxam R. and Reid N. (2003). Assessing the Effects of Transboundary Ozone Pollution between Ontario, Canada and New York, USA. Environmental Pollution (Oxford, United Kingdom), 123 (3), 403-11. • Brook, J.R., Lillyman, C.D., Shepherd, M.F. and Mamedov, A. (2002). Regional Transport and Urban Contributions to Fine Particle Concentrations in Southeastern Canada. J. Air & Waste Manage. Assoc., 52, 855-66. • Byun D.W. and Ching J.K.S (Eds.) (1999). Science Algorithms of the EPA Models-3 Community Multiscale Air Quality Model (CMAQ) Modeling System. EPA/600/R-99/030, U.S. Environmental Protection Agency, Office of Research and Development, Washington, DC 20460. • Carolina Environmental Programs (2003). Sparse Matrix Operator Kernel Emission (SMOKE) Modeling System. University of Carolina, Carolina Environmental Programs, Research Triangle Park, NC. • Draxler, R.R. and Hess, G.D. (1997). Description of the HYSPLIT_4 Modeling System. NOAA Technical Memorandum; ERL ARL-224; Silver Spring, MD, December 1997,24 pp. • Grell G.A., Duhia J. and Stauffer D. (1994). A description of the fifth-generation Penn State/NCA Mesoscale Model (MM5). NCAR Technical Note, TN-398+STR, National Center for Atmospheric Research, Boulder, CO, 138 pp. • Heidorn K.C. and Yap D. (1986). A synoptic climatology for surface ozone concentrations in Southern Ontario, 1976-1981. Atmospheric Environment, 20 (4), 695-703. • Hogrefe C., Biswas J., Lynn B., Civerolo K., Ku J.-Y., Rosenthal J., Rosenzweig C., Goldgerg R. and Kinney P.L. (2004). Simulating regional-scale ozone climatology over the eastern United States: model evaluation results. Atmospheric Environment, 38, 2627-2638. • Mukammal E.I, Neumann H.H. and Gillespie T.J. (1982). Meteorological conditions associated with ozone in Southwestern Ontario, Canada. Atmospheric Environment, 16 (9), 2095-2106. • Ontario Ministry of Environment (2001). Air Quality in Ontario, 2001 Report. Toronto, Ontario. • Yap D., Ning D.T. and Dong W. (1988). An assessment of source contributions to the ozone concentration in Southern Ontario. Atmospheric Environment, 22 (6), 1161-1168. Methods Description The system MM5/SMOKE/CMAQ was employed for the air quality analysis. The CMAQ (Byun and Ching, 1999) and MM5 (Grell et al, 1994) horizontal grid sizes were set to 36 km and 15 sigma layers were used. The domain size was located in the Northeastern part of the U.S. and Southeastern Canada (see Figure 2). CMAQ version 4.3 was used to perform gas phase chemistry simulations. Time-invariant climatological profiles for ozone and its precursors were used as boundary conditions. Version 3 of MM5 and version 2.0 of the SMOKE emissions modeling System (Carolina Environmental Programs, 2003) were used to generate, respectively, the meteorological inputs and the gridded, hourly, speciated emissions for CMAQ. The 1996 EPA National Emissions Trends (NET96) U.S. criteria inventory and the 1995 Canadian emissions inventory were used for this study. The Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT4) model (Draxler and Hess, 1997) was used to calculate 72-h back-trajectories from June 3 to July 31, 2001. Starting time of each trajectory was set at 6 AM local time and the starting height was set at 925 mb. Starting points were located in six cities, spread throughout Southern Ontario: Kingston, Egbert, Tiverton, St. Catharines, Windsor, and Toronto (see Figure 1). In order to cluster the trajectories, four transport sectors were chosen for each location (East, West, North, and South), which were qualitatively defined to denote regions with potentially different source characteristics. US observational data were obtained from 514 sites of the Aerometric Information Retrieval System (AIRS); Canadian data were obtained from 53 sites (33 located in Southern Ontario) of the National Air Pollution Surveillance (NAPS) network. Air quality model evaluation Spatial distribution The Figure 2 shows qualitative agreement but quantitative differences. Both plots show broad bands of elevated ozone across the centre of the domain, over Lakes Michigan and Erie, and along the Eastern seaboard of the United States. The model prediction, however, overestimates the absolute concentrations by up to 30 ppb in some places. Temporal distribution The agreement is good (see Figure 3), with the exception of under-predictions on day 19 (June 19) and the period from day 26 to day 30 (June 26 to 30) and over-prediction on days 50 to 54 (July 20 to 24). These periods coincided with official smog advisories, reported in the annual Air Quality Report (Ontario Ministry of Environment, 2001). Despite these discrepancies, the temporal variation is predicted quite well. The correlation coefficient for the observed and predicted 8-h daily maximum ozone concentration averaged at all 33 stations is 0.76, while the bias is 2 ppb; the slope is 0.99, and the standard error of the estimate is 8 ppb. Wind field analysis Statistical data as well as trajectory studies (Ontario Ministry of the Environment, 2001) show that ozone in Southern Ontario is strongly influenced by long range transport involving southerly and south-easterly air flows. Figure 4 shows a comparison between the predicted ozone concentration and an effective wind defined as the percentage of cells having a component of the ground level wind direction towards the North. If we only count the number of cells belonging to the closest states to Southern Ontario (from now on, we call ‘Ring’ to the group of states formed by Michigan, Ohio, Pennsylvania and New York) similar results for all U.S. territories are obtained. The correlation coefficients are around 0.5 for both scenarios. The addition of the cloud cover correction only increase slightly the correlation coefficient but the main dependence with the ozone concentration was found for the wind direction. This result qualitatively agrees with a previous study from Mukammal et al. (1982), who observed higher ozone concentration in Simcoe in presence of southerly winds. The fact that the same analysis over only the Ring’s states shows a similar correlation with the ozone concentration in Southern Ontario, suggests that the air quality in this region could be very influenced by the emissions of pollutants from the Ring’s states.