A Conceptual Model for Eight-Hour Ozone Exceedances in Houston, Texas Part I: Background Ozone Levels in Eastern Texas
Institutional Repository Document
Overview
Identity
View All
Overview
abstract
Key findings and recommendations of general interest are in boldface. Page numbers are shown in [brackets]. From Chapter 1: 1. As has been done by TCEQ, it is useful to consider the daily ozone levels in a particular area as being the sum of two quantities: the background ozone and the local contribution. (page 10) 2. The background ozone is defined as the ozone level that would be attained if there were no local anthropogenic (or unusual biogenic) emissions of ozone precursors. (page 10) 3. The local contribution is the difference between the 8-hour maximum background ozone and the 8-hour maximum actual ozone. (page 10) From Chapter 2: 4. The lowest 8-hour maximum cannot be taken to be the background ozone level in Dallas-Fort Worth (DFW) or Houston-Galveston-Brazoria (HGA) because the lowest maximum is often found in the urban core or other sites affected by local emissions. (page 13) 5. A particular limited set of stations, assumed to be measuring pristine ozone under appropriate wind conditions, is used in DFW and HGA for estimating the background ozone in those two areas. (page 13-14) 6. Individual stations within the limited set typically, but not always, record the lowest 8-hour ozone when the wind blows from the station toward the metropolitan area. (page 15-19) 7. Other regions had few monitors, so none were excluded. (page 15) 8-Hour, Part I Page 4 of 52 1/29/05 From Chapter 3: 8. Taking DFW as an example, there is year-to-year variability in the timing and magnitude of peaks of background ozone, but its interannual variability is sufficiently small that a multi-year average is an appropriate measure of typical conditions. (page 19) 9. Day-to-day variability of background ozone can be a factor of two or more, especially during late summer and early fall. (page 20) 10. There is relatively little variability of background ozone during the winter and during occasional periods in the early summer. (page 20) 11. The regular annual variation is a fundamental component of background ozone. (page 20) 12. All regions of eastern Texas have the same basic pattern of annual variation of background ozone. Background ozone is low in December but starts rising steadily in mid-January. A secondary maximum of background ozone is reached in mid-May, followed by a period with lower ozone values. The overall maximum in each region occurs in August or September, followed by a decline through the fall. (page 21-22) 13. Northeast Texas (NETX) appears to have the highest background ozone concentrations through the year. (page 22) 14. Differences in background ozone between NETX and DFW suggest that, because of sampling problems, absolute background ozone levels cannot be compared between regions with many sensors and regions with few sensors. (page 22) 15. DFW reaches a relative minimum in background ozone around the end of June and an absolute maximum in late August, while HGA reaches a relative minimum in early July and an absolute maximum in mid-September. (page 23) 16. The decline in ozone during the summer is largest in the southern regions and smallest in the northern regions. (page 23) 17. Background ozone levels as computed by the present method are suspiciously low at Beaumont-Port Arthur, suggesting that the sampling network has local source issues. (page 23) 8-Hour, Part I Page 5 of 52 1/29/05 18. The local contribution at DFW makes up less than a third of the total 8-hour ozome maximum on average days. (page 24) 19. The local contributions at DFW and HGA peak in July and August. (page 24-25) 20. The local contribution at HGA is generally a greater percentage of the total ozone than at DFW and reaches an average value of 0.035 ppmv in August. (page 25) 21. The local contribution at HGA averages 0.010 ppmv greater than at DFW, while the background at HGA averages 0.010 ppmv less than at DFW, leading to similar mean 8-h ozone levels. (page 25) 22. Background ozone levels at HGA and DFW correlate most strongly with the wind direction (and component from the north) on the previous day, with slightly lower but still highly significant correlations on the same day and two days previous. (page 26) 23. Also contributing to background ozone are weak winds at HGA and a lack of precipitation at DFW. (page 27) 24. The local contribution at DFW and HGA is positively correlated with high temperatures, low wind speeds, and a lack of precipitation. (page 27) 25. On days without precipitation, the average 8-h maximum ozone in HGA in early September exceeds the 8-h standard. Ozone in DFW comes close. (page 28) From Chapter 4: 26. Principal Component Analysis (PCA) is useful for representing the large-scale winds, which may be expected to strongly control background ozone, as a small number of continuously-varying patterns. (page 30) 27. Mean ozone-season winds are from the southeast in the HGA area and from the south in the DFW area, rotating clockwise around a center of high pressure in the southeastern United States. (page 31) 28. The leading principal component (PC1), which explains the greatest amount of variability in the wind pattern, represents winds from the southwest (if positive) or northeast (if negative). It therefore is an indicator of the presence (or absence, if positive) of transport from the central and eastern United States. (page 31) 8-Hour, Part I Page 6 of 52 1/29/05 29. The second principal component (PC2) is positive with winds from the northwest. (page 31-32) 30. The total wind on any given day is the sum of the mean wind and the daily amplitudes of the various principal components. (page 33) From Chapter 5: 31. The principal component most strongly correlated with background ozone is PC1, followed by PC5 and PC2. (page 37) 32. The correlations are strongest with the wind pattern one day before the ozone event. (page 37) 33. By itself, PC1 explains background ozone variations of about 0.017 ppmv at DFW and over 0.020 at HGA. (page 38) 34. The mean value of PC1 declines steadily from s
