1. Introduction
Ground-level ozone, associated with manufacturing and automotive combustion processes, is a serious health hazard in many locations where vicinity to ozone sources combined with local atmospheric conditions and geographic features can lead to an unhealthful buildup of ozone. Under the Clean Air Act, the U.S. Environmental Protection Agency (EPA) is required to set National Ambient Air Quality Standards (NAAQS) for ozone. Currently, the EPA defines a “high ozone event” as any 1-h episode in which ozone levels exceed 0.12 ppm or, alternatively, as any 8-h period for which the average ozone level exceeds 0.075 ppm (as of 27 May 2008). Health and government professionals typically issue high ozone warnings when ozone levels (i.e., mixing ratios) are poised to exceed this standard, based on current ozone levels and local atmospheric conditions. In most cases, high ozone warnings are issued within 24 h of the scheduled ozone event. Considering the human health cost involved, it is clearly desirable to provide longer-term warnings to the public if possible.
Many studies have been conducted to identify the local atmospheric conditions and topographical features that favor high ozone events in certain areas. Some recent research includes studying hourly wind patterns associated with sea and gulf breezes, thermal-driven flows, thunderstorm outflows, and stagnant pressure systems in Houston, Texas (Darby 2005); looking at the interplay of local onshore breezes with synoptic-scale pressure systems in the San Luis Obispo County and southwestern San Joaquin Valley region (Niccum et al. 1995); identifying the role of vertical mixing and stratified boundary layers in ozone exceedance in the northeastern United States (Zhang et al. 1998; Zhang and Rao 1999); relating high ozone events to synoptic high pressure, stratospheric processes, and local diurnal wind forcing at high elevation in North Carolina (Aneja et al. 1991); and assessing the role of topography on ozone transport in the Great Smoky Mountains National Park (Mueller 1994). Certainly, identifying the local factors that influence ozone levels is of primary importance. However, the prediction of these factors relies on small temporal-scale and spatial mesoscale modeling of both atmospheric and chemical processes. These models are not always reliable for ozone forecasts because of the complexities of local terrain (Barna et al. 2000), and they are limited in their forecast lead times. If larger-scale atmospheric factors can be linked to high ozone events, ozone warning lead times could potentially be increased by relying on routine medium-range weather forecasts.
To that end, this study attempts to link large-scale and regional meteorological factors to ground ozone levels, since those meteorological factors can be forecast relatively accurately up to 72 h in advance. Of particular interest is vertical velocity near the top of the planetary boundary layer, which may be expected to influence surface ozone by changing both ventilation and depth of the boundary layer. Midlevel (500 hPa) geopotential height is also examined, as it provides a good indication of the large-scale circulation and is one of the best forecast fields in medium-range weather prediction (e.g., Kalnay et al. 1998). The analysis is primarily performed relative to a single ozone monitoring station, to assess the potential for local forecasting. The Pasadena ozone monitoring station is selected as the primary analysis site for Southern California because it is close to sources of ozone (manufacturing; automobile exhaust) and has a history of high ozone events. In addition, daily data are available for the full 8 years, with no more than one consecutive day missing at any point during the time period. The correlation between daily ozone variability in Pasadena and other stations in California is calculated to quantitatively assess the representativeness of the station.
2. Data and methods
a. Ozone data
July–September 1994–2001 hourly ozone data for Pasadena, California, are gathered using the EPA’s Air Quality System (AQS) database available through the EPA’s Technology Transfer Network (TTN). The monitoring site chosen is located at 34.1°N, 118.1°W and carries EPA site designation 060372005. The hourly data are collected using ultraviolet spectroscopy in accordance with the EPA’s reference method EQOA-0992–087 or equivalent Model 400 method. In the case of a missing day, the average of the previous and next day’s ozone level is used. Since approximately 80% of daily high ozone levels occur from 2000 to 2200 UTC, this period’s ozone readings are averaged and used as the daily ozone level in parts per billion. For these data, an upper threshold of 120 ppb corresponds to the highest 10% of the days; a corresponding low threshold of 43 ppb corresponds to the lowest 10% of the days. These thresholds result in 76 high days and 77 low days out of a pool of 736 days. The composited ozone data are then compared with various meteorological data for those same days.
The daily ozone levels used for Pasadena are shown as a time series in Fig. 1. The high and low thresholds are shown as light gray lines in the figure. There is a modest downward trend in the ozone levels throughout the 8-yr period, likely due to California’s strict emission standards implemented during this time period. However, this slight downward trend does not affect the results of this study, as similar results are found using both the first four years and last four years of the period.
Daily surface ozone for Pasadena, July–September, 1994–2001. The thresholds for high- and low-ozone days are shown as light gray lines.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Daily surface ozone for Pasadena, July–September, 1994–2001. The thresholds for high- and low-ozone days are shown as light gray lines.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Daily surface ozone for Pasadena, July–September, 1994–2001. The thresholds for high- and low-ozone days are shown as light gray lines.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Pasadena’s average daily ozone level is also correlated to the average daily ozone levels of 56 other ozone-reporting sites in California. The 56 sites are selected based on the same criteria used to select Pasadena (no more than one consecutive day at a time missing throughout the 736-day period); however, the daily ozone level used is the 24-h average for each of the 57 sites.
b. Meteorological data
National Centers for Environmental Prediction–National Center for Atmospheric Research reanalysis data (Kalnay et al. 1996) are used to determine daily values of various meteorological data, in particular vertical velocity at 700 hPa and geopotential height at 500 hPa, for the same 736-day dataset as ozone (July–September 1994–2001). The 700-hPa level is chosen for vertical velocity since it is generally above the surface boundary layer and likely to capture overall vertical movements. The 500-hPa geopotential height level is chosen to best capture large-scale synoptic features. The dataset domain spans the latitude range 15°–50°N and the longitude range 160°–80°W, with grid resolution set at 2.5°. Composites of these meteorological variables are created for the high- and low-ozone days, as well as for the full 736 days.
c. Significance
A random resampling Monte Carlo approach is used to determine field significance. Specifically, 1000 composite sets of the meteorological variables are produced by random sampling from the 736-day pool; the high- and low-ozone day composites for each meteorological variable are then compared with the 1000 random composites of these variables. Both local significance (at a single grid point) and global significance (to eliminate random local significance) are assessed. For this study, a significance level of 99% is used.
3. Results
a. Regional ozone correlation
The local ozone data at Pasadena are correlated to other ozone-reporting sites in California, and the results are shown in Fig. 2. Pasadena is shown as a black circle. Red circles indicate positive correlation; blue circles indicate negative correlation; and in both cases darker shades indicate higher correlations. Correlations higher than 0.5 are outlined in black. A cluster of 14 high-correlation stations surround Pasadena, with 13 of the 14 correlations statistically significant at the 0.05 level. This high correlation of ozone data between sites near Pasadena in Southern California (roughly within 200 km) demonstrates Pasadena’s representativeness for the region. Similar results are found by correlating only high-ozone days and low-ozone days (not shown). While many of the ozone-generating processes and sources are local, the analysis suggests that there are also factors that influence the variability at regional scales.
Correlation of other California ozone stations to Pasadena ozone. Red shades denote positive correlation; blue shades denote negative correlation. Correlations >0.5 are outlined in black. Stations are chosen based on consistent data availability for the 1994–2001 period.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Correlation of other California ozone stations to Pasadena ozone. Red shades denote positive correlation; blue shades denote negative correlation. Correlations >0.5 are outlined in black. Stations are chosen based on consistent data availability for the 1994–2001 period.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Correlation of other California ozone stations to Pasadena ozone. Red shades denote positive correlation; blue shades denote negative correlation. Correlations >0.5 are outlined in black. Stations are chosen based on consistent data availability for the 1994–2001 period.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
b. Regional influences: Vertical velocity
The relationship between Pasadena ozone and 700-hPa vertical velocity is shown in Fig. 3: Fig. 3a shows the average vertical velocity during the period of interest (July–September 1994–2001); Fig. 3b shows the average vertical velocity difference between high- and low-ozone days with differences significant to the 99% level shaded; and Figs. 3c and 3d show the average vertical velocity during low- and high-ozone days, respectively. For reference, Pasadena is indicated in the figures by a black circle. We have verified that the changes in vertical velocity extend down to 1000 hPa.
Association of 700-hPa vertical velocity (Pa s−1) with high- and low-ozone days: (a) the average vertical velocity over the full period, (b) the difference between the high and low averages, with areas of local 99% significance shaded, (c) the average during low-ozone days, and (d) the average during high-ozone days. Contour interval is 0.02 Pa s−1 in (a), (c), and (d) and 0.01 Pa s−1 in (b).
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Association of 700-hPa vertical velocity (Pa s−1) with high- and low-ozone days: (a) the average vertical velocity over the full period, (b) the difference between the high and low averages, with areas of local 99% significance shaded, (c) the average during low-ozone days, and (d) the average during high-ozone days. Contour interval is 0.02 Pa s−1 in (a), (c), and (d) and 0.01 Pa s−1 in (b).
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
Association of 700-hPa vertical velocity (Pa s−1) with high- and low-ozone days: (a) the average vertical velocity over the full period, (b) the difference between the high and low averages, with areas of local 99% significance shaded, (c) the average during low-ozone days, and (d) the average during high-ozone days. Contour interval is 0.02 Pa s−1 in (a), (c), and (d) and 0.01 Pa s−1 in (b).
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
On average, during the summer July–September period considered here, there is subsidence over the eastern Pacific, while at the same time an area of rising motion exists over the generally elevated terrain of the western United States and northwestern Mexico. On average for July–September, Pasadena is in an area of slightly sinking motion at 700 hPa (0.01 Pa s−1). However, during high-ozone days, there is a shifting of downward motion farther inland in Southern California (resulting in Pasadena being in an area of 0.04 Pa s−1 downward motion), although the magnitudes of upward and downward motion remain the same (Fig. 3d). During low-ozone events, there is a shifting of the boundary between upward and downward motion to an offshore position in Southern California (Fig. 3c), resulting in upward motion over Pasadena (−0.03 Pa s−1). In addition, the magnitude of upward vertical velocity centered over Nevada is higher than average for low-ozone events. As seen in Fig. 3b, the largest difference between the two composites for vertical motion at 700 hPa occurs over Southern California. These results show that ozone variability in Southern California has a statistically significant link to vertical motion at 700 hPa, with a distinct contrast between upward motion during low-ozone events and downward motion during high-ozone events. The magnitude of the difference between the two composites (Fig. 3b) is comparable to the largest values of mean ascent and descent over the entire region (Fig. 3a).
c. Large-scale influences: Geopotential height
The large-scale geopotential height field also shows a close association with ozone levels. Figure 4 shows the average 500-hPa geopotential heights for July–September 1994–2001, the height composites during high-ozone and low-ozone events during this time period in Pasadena, and the difference between these two composites, shaded at the 99% significance level. Of particular interest is the relative strength and location of seasonal circulation patterns. Figure 4a shows the North American anticyclone (NAA) in its favored summer position over Texas and New Mexico (Adams and Comrie 1997). The subtropical North Pacific anticyclone (PA) is in place in the lower troposphere in the northern Pacific (Rodwell and Hoskins 2001). During high-ozone days in Pasadena, there is a northwestward shift in the location of the NAA, as well as an increase in its magnitude (Fig. 4d). At the same time the PA shifts to the southwest. On the other hand, during low-ozone events in Pasadena, there is a weakening and southeastward shifting of the NAA. A ridging concurrently occurs in the mid-Pacific, linked to a northeast shift in the PA, causing an upper-level trough just offshore of southern California (Fig. 4c). The 99% statistically significant difference in heights between low- and high-ozone events in Pasadena occurs in a broad area over the western United States, with the strongest differences over California and Nevada (Fig. 4b), and these changes in the composite heights determine whether Southern California is under cyclonic or anticyclonic flow aloft. Surface ozone variability in Southern California, therefore, is linked with significant differences in the midlevel height field: changes to the position and strength of the NAA, the presence or absence of a trough off the West Coast, and a change between cyclonic and anticyclonic flow directly aloft.
As in Fig. 3, but for 500-hPa geopotential height.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
As in Fig. 3, but for 500-hPa geopotential height.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
As in Fig. 3, but for 500-hPa geopotential height.
Citation: Journal of Applied Meteorology and Climatology 50, 4; 10.1175/2010JAMC2605.1
4. Summary and discussion
Analysis of the links between daily ozone in Southern California and atmospheric circulation for an 8-yr period shows that daily ozone variability is linked to regional vertical velocity and to large-scale changes in horizontal circulation, including striking changes in the position and strength of the North American anticyclone.
The link to vertical velocity is consistent with the strong influence of vertical velocity on the ventilation and depth of the boundary layer (e.g., Seinfeld 1986; Stull 1988; National Research Council 1991; Vallero 2008), which in turn affects the accumulation of ozone and its precursors near the surface. The link to continental-scale circulation features remains to be fully explored, although it is consistent with the close relationship between the anticyclone and vertical velocity (e.g., Rodwell and Hoskins 2001).
While trends in air quality affect the stability of the time series—indeed, a downward trend in ozone over the period can be seen in Fig. 1—we have verified that similar results were obtained for both the first four and last four years of the period when analyzed separately, so that the physical processes captured here appear relatively insensitive to the trend.
In addition to the scientific interest in understanding the links between large and regional-scale atmospheric circulation and local ozone, the link to large-scale changes in a variable (500-hPa geopotential height) that is one of the best forecast fields in numerical weather prediction suggests the possibility of important practical applications of current medium-range weather forecasts for medium-range forecasting of ozone. We are currently investigating this potential. A simple measure, however, of the potential for using the midlevel circulation to predict ozone variability can be established by looking at the areal average of daily geopotential height over Pasadena (using the nine 2.5° × 2.5° grid boxes centered over the station) and assessing the average daily ozone at Pasadena for the highest and lowest 10% of the height averages. The highest 10% of the height averages corresponds to an average ozone level at Pasadena of 99 ppb, whereas the lowest 10% of the height averages correspond to an ozone level of 48 ppb, suggesting that the midlevel circulation provides considerable information about the surface ozone. Exploring the potential of more sophisticated measures of circulation in forecasting ozone is the subject of ongoing work.
Acknowledgments
This paper is based in part on the master’s thesis work of Vianney Lopez, University of Massachusetts Lowell, 2007, and was supported by NSF Grant 0621237 and NOAA Grant NA08OAR4310592.
REFERENCES
Adams, D. K., and A. C. Comrie, 1997: The North American monsoon. Bull. Amer. Meteor. Soc., 78, 2197–2213.
Aneja, V. P., S. Businger, Z. Li, C. S. Claiborn, and A. Murthy, 1991: Ozone climatology at high elevations in the southern Appalachians. J. Geophys. Res., 96, 1007–1021.
Barna, M., B. Lamb, S. O’Neill, H. Westberg, C. Figueroa-Kaminsky, S. Otterson, C. Bowman, and J. DeMay, 2000: Modeling ozone formation and transport in the Cascadia region of the Pacific Northwest. J. Appl. Meteor., 39, 349–366.
Darby, L. S., 2005: Cluster analysis of surface winds in Houston, Texas, and the impact of wind patterns on ozone. J. Appl. Meteor., 44, 1788–1806.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Kalnay, E., S. Lord, and R. McPherson, 1998: Maturity of operational numerical weather prediction: Medium range. Bull. Amer. Meteor. Soc., 79, 2753–2769.
Mueller, S. F., 1994: Characterization of ambient ozone levels in the Great Smoky Mountains National Park. J. Appl. Meteor., 33, 465–472.
National Research Council, 1991: Rethinking the Ozone Problem in Urban and Urban and Regional Air Pollution. National Academy Press, 524 pp.
Niccum, E. M., D. E. Lehrman, and W. R. Knuth, 1995: The influence of meteorology on the air quality in the San Luis Obispo County–southwestern San Joaquin Valley region for 3–6 August 1990. J. Appl. Meteor., 34, 1834–1847.
Rodwell, M. J., and B. J. Hoskins, 2001: Subtropical anticyclones and summer monsoons. J. Climate, 14, 3192–3211.
Seinfeld, J., 1986: Atmospheric Chemistry and Physics of Air Pollution. Wiley-Interscience, 786 pp.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.
Vallero, D., 2008: Fundamentals of Air Pollution. 4th ed. Academic Press, 942 pp.
Zhang, J., and S. T. Rao, 1999: The role of vertical mixing in the temporal evolution of ground-level ozone concentrations. J. Appl. Meteor., 38, 1674–1691.
Zhang, J., S. T. Rao, and S. M. Daggupaty, 1998: Meteorological processes and ozone exceedances in the northeastern United States during the 12–16 July 1995 episode. J. Appl. Meteor., 37, 776–789.
