[NOTE: Figures are available from WALCOFF
(703) 267-3594]


Final Report
CLASSIFICATION OF OZONE EPISODES FOR FOUR SOUTHERN CITIES ACCORDING TO TRANSPORT CHARACTERISTICS

SYSAPP-97/03
March 1997
Prepared for John Jansen
Southern Company Services
600 North 18th Street
Birmingham, AL 35203

Prepared by A. Belle Hudischewskyj and Sharon G. Douglas
Systems Applications International, Inc.
101 Lucas Valley Road
San Rafael, California 94903
Copyright 1997 by Systems Applications International, Inc.


Contents
Executive Summary

  1. INTRODUCTION
  2. OVERVIEW OF THE EPISODE CLASSIFICATION METHODOLOGY
  3. CHARACTERIZATION OF OZONE EPISODES
  4. SUMMARY OF FINDINGS

1 INTRODUCTION 

This report summarizes the results of an ozone episode classification analysis that was performed for the cities of Atlanta, Birmingham, Charlotte, and Nashville using the Classification and Regression Tree (CART) analysis technique as the primary classification tool. The overall objective of the analysis was to characterize ozone exceedance days for the four urban areas as either predominantly "home grown" (resulting from locally emitted pollutants) or resulting from pollutant transport. Following classification of the exceedance days, the potential influence of transport with respect to attainment of the federal ozone standard was also examined.

In conducting this analysis, we assumed that all ozone episodes have both a home-grown and a transport component. The problem then became one of determining the relative contribution from transport. In order to make this determination, the CART analysis technique was applied separately for each of the four urban areas. All ozone season days for the years 1986 through 1995 were classified based on maximum observed ozone concentration and according to key meteorological parameters.

The CART analysis technique is a statistical analysis tool that partitions a dataset into discrete subgroups based on the value of a user-defined classification variable (e.g., maximum ozone concentration). The technique is described by Brieman et al. (1984) and has been applied for a wide variety of purposes. The technique is designed to segregate objects, or in the case of air quality analysis, days with different values of a classification variable into different bins or terminal nodes. The CART technique accomplishes this task through the growth of a binary decision tree, comprised of a progression of binary splits on the values of the other input variables. Each split is chosen such that the segregation of different values of the classification variable is improved. The resulting tree has multiple branches, of various complexity, each of which represents a variable path to a series of values of the classification variable.

 

For this analysis, CART was applied to the classification of ozone episode days. The dependent or classification variable was specified to be maximum ozone concentration (at any monitoring site within the urban area). Four classification bins were defined corresponding to maximum ozone concentrations: (1) less than 85 ppb, (2) greater than or equal to 85 ppb but less than 105 ppb, (3) greater than or equal to 105 but less than 125 ppb, and (4) greater than or equal to 125 ppb. Exceedance days were, thus, classified separately.

The independent meteorological variables used for the analysis include: maximum surface temperature, average surface wind speed for three time periods during the day (0700-1000, 1000-1300, and 1300-1600 LST), relative humidity at midday (1200 LST), cloud cover at three times during the day (1000, 1200, and 1400 LST), difference in temperature between the 850 mb level and the surface (morning), 850 and 700 mb wind speed (average of the morning and evening measured values), and 850 and 700 mb geopotential height, an indicator of synoptic-scale pressure conditions (average of the morning and evening values). Wind direction was not included, as there is no well defined relationship between wind direction and transport into the areas under study.

Additional air quality variables that were included in the CART analysis are: time of maximum ozone concentration (four time bins were used to represent this variable), average maximum ozone concentration (averaged over all sites within the area under study), and previous days maximum ozone concentration. The presence of a secondary morning ozone peak and the average increase in ozone during the morning hours (averaged over all sites) were also examined, although not included in the CART analysis.

Following application of the CART analysis technique, the characteristics of days within each terminal bin corresponding to an ozone exceedance bin were examined and the days within each bin were characterized as either predominantly home-grown or resulting from pollutant transport, as described below. A number of factors were considered in this determination.

Indicators of home-grown episodes include high temperatures, low wind speeds near the surface and aloft, clear skies, high 850 mb temperatures, high 850 and 700 mb heights, and peak ozone between 1400 and 1600 EST. Isolated high ozone (high peak but a lower average ozone concentration) is also an indicator of a local event. A secondary ozone peak that occurs during the morning or a rapid rise in ozone during the morning hours coupled with high ozone on the previous day and light and variable winds aloft suggest recirculation.

Indicators of transport-induced episodes include non-ozone-conducive local meteorological conditions, high wind speeds near the surface and/or aloft, early or late ozone peaks, and a rapid increase in ozone concentration during the morning hours coupled with high wind speeds aloft.

The characteristics of the days within each terminal bin containing one or more exceedance days were examined and compared with respect to these indicators and the relative contribution from transport was estimated. It is important to note that transport was considered only from the perspective of transport into the urban areas under study; pollutant transport and fate of pollutant out of the areas was not considered in this analysis.

Repeating sequences of classification types were also examined to aid in the identification of multi-day pollutant recirculation and transport episodes.

The results of this analysis are presented in the remainder of this report. An overview of the classification methodology is presented in Section 2. The CART results for Atlanta, Birmingham, Charlotte, and Nashville, including a discussion of classification accuracy, are presented in Section 3. The characterization of ozone episode days within each exceedance bin as home-grown or transport dominated is also provided in Section 3. A summary and concluding remarks are presented in Section 4.


2 OVERVIEW OF THE EPISODE CLASSIFICATION METHODOLOGY

CART ANALYSIS

CART (Classification and Regression Tree) analysis is a binary splitting method which partitions a dataset into discrete subgroups based on the value of a user-defined classification variable. The remaining variables in the database are selected to provide a basis for segregation of the data among different values of the classification variable. Thus a causal relationship between these variables and the classification variable is assumed. Consequently, it is necessary to construct a database of independent variables upon which the classification or dependent variable is likely to depend.

The classification variable in this case is a representation of daily maximum ozone concentration. The classification technique depends upon this variable being discrete, although the independent variables may be continuous. The classification variable employed in the present analysis was based on the value of the maximum observed ozone concentration for a particular urban area of interest. Specifically, this variable was assigned a value of 1 to 4, corresponding to a maximum ozone concentration of less than 85, 85 to less than 105, 105 to less than 125, or greater than or equal to 125 ppb. It was assigned a value for each day, where the maximum was chosen from observations from a network of air quality stations (specified below).

Databases consisting of meteorological variables and other air quality parameters were constructed for four urban areas: Atlanta, Georgia; Birmingham, Alabama; Charlotte, North Carolina; and Nashville, Tennessee. These data constituted the independent variables for the application of CART. Each of these databases consisted of meteorological variables characterizing the prevailing conditions in the local area for each day as well as of observed ozone concentrations for the local area on the previous day. (The ozone concentrations on the previous day were included to enable CART to distinguish same-day local ozone formation from recirculation or transport of ozone.) The variables constructed for each of the four urban areas were drawn from data collected from air quality monitoring stations, surface meteorological monitoring stations, and upper-air meteorological monitoring stations. The exact variables and the monitoring stations employed for each database are specified below.

Air Quality Variables

The air quality variables used in the CART analysis for each urban area are defined in Table 2-1.

Meteorological Variables

The surface meteorological variables used in the CART analysis for each urban area are defined in Table 2-2, and the upper-air meteorological variables used in the CART analysis are defined in Table 2-3.

 

 

TABLE 2-1. Air quality variables included in the CART analysis.

Variable Name

 

Description

ymxo3all

Maximum ozone concentration observed within the urban area on the previous day.

ratmaxav

The percent difference between the maximum and the average maximum ozone concentration observed within the urban area on the present day.

bintim

A time "bin" corresponding to the time of occurrence of the maximum ozone concentration. The value equals 1 if the maximum concentration occurs at or before 1100 LST, it equals 2 if it occurs between 1200 LST and 1600 LST, and it equals 3 if occurs after 1600 LST.

o3class

The classification variable: a value of 1,2,3, or 4 depending on whether the maximum ozone concentration over all sites in the urban area on the present day was <85, [85,105), [105,125), or 125 ppb, respectively.

 

 

TABLE 2-2. Surface meteorological variables included in the CART analysis.

Variable Name

Description

ws710

Average surface wind speed (m/s) from 0700 to 1000 LST

ws1013

Average surface wind speed (m/s) from 1000 to 1300 LST.

ws1316

Average surface wind speed (m/s) from 1300 to 1600 LST.

tmax

Maximum surface temperature ( C) for the present day.

rh12

Surface relative humidity at noon.

ccl10

A value of 0 or 1, indicating that the opaque cloud cover at 1000 LST was 50 percent or >50 percent, respectively.

ccl12

A value of 0 or 1, indicating that the opaque cloud cover at 1200 LST was 50 percent or >50 percent, respectively.

ccl14

A value of 0 or 1, indicating that the opaque cloud cover at 1400 LST was 50 percent or >50 percent, respectively.

 

 

TABLE 2-3. Upper-air meteorological variables included in the CART analysis.

Variable Name

Description

avgws(pmb)

The average of the morning and afternoon wind speeds (ms-1) at the (pmb)a level determined from the twice daily soundings.

avght(pmb)

The average of the morning and afternoon geopotential heights (m) at the (pmb) level determined from the twice daily soundings

upmsfcam

The difference between the morning temperature at the 850 mb level and the surface temperature ( C) at the same time.

a (pmb) is either 700 mb or 850 mb.

 

Summary of Monitoring Sites

Separate datasets were created for the Atlanta, Charlotte, Birmingham, and Nashville urban areas. The data set for each of these cities consisted of data from either four (Atlanta) or three (Charlotte, Birmingham, Nashville) air quality monitoring stations, one upper-air meteorological station (note that in 1995, the upper-air sites for Atlanta and Birmingham were replaced by new sites), and one surface meteorological monitoring site. Each of the above variables were tabulated for each day from April through October for the years 1986 through 1995, inclusive. The station identifiers and locations are summarized in Table 2-4. These sites were selected based on an examination of data completeness for the ten-year period.

 

TABLE 2-4. Meteorological and air quality monitoring sites included in the CART analysis.

Primary City

Station Type

Mnemonic ID

Station IDa

Atlanta, GA

Air Quality

South DeKalb

130890002

 

Air Quality

Conyers-Monastery

132470001

 

Air Quality

MLK Marta Station

131210053

 

Air Quality

Sweetwater

130970002

 

Surface Met.

Atlanta, GA NWS

13874

 

Upper Met.

Athens, GA NWS (1986-1994)

13873

 

 

Atlanta, GA NWS (for 1995)

53819

Charlotte, NC

Air Quality

Plaza Road and Lakedell

371190034

 

Air Quality

Westinghouse Blvd.

371191005

 

Air Quality

Mecklenburg Cab Co.

371191009

 

Surface Met.

Charlotte/Douglas Airport

13881

 

Upper Met.

Greensboro, NC NWS

13723

Birmingham, AL

Air Quality

Fairfield Fire Dept.

010731003

 

Air Quality

Pinson High School

010735002

 

Air Quality

Tarrant Elementary School

010736002

 

Surface Met.

Birmingham Municipal Airport

13876

 

Upper Met.

Centerville, AL NWS (1986-1995)

03881

 

 

Birmingham, AL NWS (1995)

53823

Nashville, TN

Air Quality

Trinity Lane

470370011

 

Air Quality

Percy Priest Visitors Center

470370026

 

Air Quality

Rockland Rec. Area

471650007

 

Surface Met.

Nashville Metro Airport

13897

 

Upper Met.

Nashville Metro Airport

13897

a For air quality stations, the AIRS identifier is given, while for surface and upper air meteorological stations, the WBAN identifier is given.

 

 

CART Parameters

The method employed by the CART software for minimizing misclassification and optimizing the robustness of the derived classification scheme depends on how misclassification and robustness are measured. These measures are a function of a series of user-definable parameters, the settings for which are listed in Table 2-5. A complete description of these parameters can be found in the on-line documentation for the software, as well as the text by Brieman et al. (1984). The misclassification "costs" with which the errors are calculated for the Atlanta, Birmingham, Charlotte, and Nashville datasets are shown in Table 2-6.

EXAMINATION OF EXCEEDANCE BIN CHARACTERISTICS

Following application of the analysis technique, the CART results were examined to ensure that the parameters and parameter values corresponding to the binary splits and resulting in the assignment of days to the terminal bins are physically reasonable. The amount and degree of misclassification were also examined.

 

 

 

TABLE 2-5. CART input/controlling parameters.

Parameter

Setting

Construction rule

Ordered twoing

Estimation method

Four-fold cross validation

Class priors

User specified

Tree selection

0.0 standard error rule.

Linear combinations

Not allowed

Maximum number of surrogates

15

Misclassification costs

In general, a linearly increasing series giving greater cost for greater misclassification. In addition, there is an increased cost across the non-exceedance/exceedance transition.

 

 

TABLE 2-6. Misclassification costs for Atlanta, Birmingham, Charlotte, and Nashville CART applications.

 

 

Cost if

Classified As

 

 

True Class

1 : <85

2 : [85,105)

3 : [105,125)

4 : 125

1 : <85

0.0

1.0

2.0

3.0

2 : [85,105)

1.0

0.0

1.0

2.0

3 : [105,125)

2.0

1.0

0.0

1.0

4 : 125

3.0

2.0

1.0

0.0

 

The characteristics of each terminal bin containing one or more exceedance days were then reviewed as follows:

A detailed analysis of the parameter and parameter values associated with the binary splits leading to the terminal bin was performed (this provided information on the distinguishing features of days within the bin, compared to days in other bins)

Average values of each CART input variable were calculated for each bin (using all days assigned to the bin as well as only those that were correctly classified as belonging to the category represented by the bin) and compared

In this manner, a relative assessment of the local meteorological characteristics with respect to ozone conducive conditions (as described in the previous section) was obtained.

CLASSIFICATION OF EXCEEDANCE DAYS ACCORDING TO TRANSPORT POTENTIAL

Considering the overall characteristics of the days within each bin, the potential contribution from ozone transport was then estimated. Primary consideration was given to wind speed. The average wind speeds (both near the surface and aloft, with the emphasis on wind speeds aloft) were classified as low, moderate, or high based upon the distribution of observed values for the study period. Days within bins characterized by moderate to high wind speeds were classified as having a moderate to high potential for a contribution from transport. This classification was adjusted according to whether the remaining meteorological descriptors were consistent with local ozone conducive conditions (for example, high temperatures, clear skies). Days within bins characterized by low wind speeds and other features indicative of local ozone formation were classified as home grown.

Many ozone exceedance days that occurred during the study period were part of multi-day ozone episodes (during which exceedances or values close to the exceedance value were observed on consecutive days). To accommodate this feature in characterizing the exceedance days, the potential influence of day-to-day carryover or recirculation of ozone on the exceedance days was also estimated as part of this analysis. The previous days maximum ozone concentration was classified as low, moderate, or high; these values corresponded to Categories 1 and 2 (low), 3 (moderate) and 4 (high). Wind speeds aloft were also classified as described earlier. Days within bins characterized by low wind speeds and high ozone concentrations on the previous day were classified as having a high potential for recirculation. Days within bins characterized by high wind speeds and low ozone concentrations on the previous day were classified as having a low potential for recirculation. Other days were assigned to intermediate classification categories, based on the values of these parameters.

Finally, the classification tree was examined with respect to recurring sequences of bins. In particular, the incidence of transport-dominated days followed by home grown exceedance days was considered in the final designation of the exceedance bins as home grown or transport dominated.

3 CHARACTERIZATION OF OZONE EPISODES

The characterization of ozone episodes for each of the four urban areas included application of the CART analysis technique for classifying the episodes according to maximum ozone concentration and key meteorological parameters or distinguishing features, examination of the days within each exceedance bin or category with respect to transport conducive characteristics, and qualitative assessment of the potential for contribution from transport. The analysis results for each urban area are discussed separately in the following sections.

ATLANTA

The results of the CART analysis for the Atlanta area are discussed in this section. This discussion is followed by a detailed analysis of the meteorological and transport characteristics of the days within the exceedance bins. Days on which exceedances of the federal ozone standard were recorded that were not classified correctly into exceedance bins are examined separately.

As indicated in the previous section, key classification parameters include wind speeds (near the surface and aloft) and the previous days maximum ozone concentration. In the following summary for the Atlanta area, the descriptors low, moderate, and high for surface wind speeds refer to those that are less than 3 ms-1, between 3 and 5 ms-1, and greater than 5 ms-1, respectively. For the upper levels, they refer to wind speeds that are less than 4 ms-1, between 4 and 9 ms-1, and greater than 9 ms-1, respectively. For ozone, the descriptors refer to concentrations that are less than 85 ppb, between 85 and 105 ppb, and greater than 125 ppb, respectively.

CART Results

The classification tree for the Atlanta urban area consists of 32 terminal bins (Figure 3-1). Of these, 6 bins are designated exceedance bins and contain mostly days for which the maximum ozone concentration at any site within the Atlanta urban area exceeded the current federal ozone standard of 124 ppb. The classification accuracy of the CART tree is depicted in Table 3-1. Good classification is achieved within most categories and, in most cases, days that are misclassified are put into an adjacent category. During the ten-year period included in this analysis, there are 107 exceedance days. Ninety-eight of these are properly classified as having maximum ozone concentrations that are greater than or equal to 125 ppb (Category 4). Nine exceedance days are misclassified; eight of these as having a maximum ozone concentration greater than or equal to 105 ppb but less than 125 ppb (Category 3) and one day is misclassified as having a maximum ozone concentration that is greater than or equal to 85 ppb but less than 105 ppb (Category 2). The nine exceedance days that are misclassified are placed in seven different terminal bins. There is greater misclassification of non-exceedance days into the exceedance category; a total of 55 days are misclassified as exceedances (forty percent of these have ozone maxima that are between 105 and 125 ppb; the remainder are associated with lower ozone concentrations). It is interesting to note, that, of the days that are incorrectly classified as exceedance days, approximately 30 percent are weekend days. Some misclassification of weekend days into higher ozone categories is expected, due to lower emissions on weekends.

Certain variables were identified by the CART program as being of primary importance in constructing the binary classification tree for the Atlanta area. These include yesterdays maximum ozone concentration, average surface wind speed between 1000 and 1300 LST, average wind speed at the 850 mb level, average surface wind speed between 1300 and 1600 LST, average surface wind speed between 0700 and 1000 LST, average wind speed at the 700 mb level, relative humidity at noon, and average height of the 700 mb surface.

Review of these important parameters and the classification tree (the parameter splits and the average values of the input parameters associated with each terminal bin) indicates that the relationship between the meteorological variables and the ozone concentrations depicted by the CART tree is a physically meaningful one. For example, considering all bins within each of the four classification categories 1 through 4, corresponding to increasing maximum ozone concentration, the average daily maximum temperature increases with each higher category. Conversely, relative humidity, cloud cover, and surface wind speed decrease with each higher category. These findings conform to known relationships between high ozone concentrations and high temperatures, clear skies, and low wind speeds.

Exceedance Bin Characteristics

Bin 2

Although Bin 2 is primarily a Category 3 bin (with maximum ozone concentrations between 105 and 125 ppb), one exceedance day was placed in this bin. Days within this bin are characterized by a maximum surface temperature between 27.5 and 31.4 C (84.3 and 88.5 F), low relative humidity at midday (less than 53.5 percent), and very low surface wind speeds between 1000 and 1300 LST (less than 1.54 ms-1). Days within this bin are also characterized by a peak ozone concentration that is within approximately 18 percent of the average maximum concentration, that is, the maximum ozone concentration for most sites is similar in magnitude. Average wind speeds aloft are also very low. The low wind speeds suggest a dominant home-grown component. Relative high maximum ozone on the day prior to each day within the bin indicates possible recirculation. Thus, the exceedance day within this bin is classified as home-grown (and possibly influenced by day-to-day carryover of ozone within the Atlanta area).

Bin 7

The first twelve terminal bins of the Atlanta CART tree contain those days with relatively low daily maximum surface temperatures (the first branching separates those days with a maximum temperature greater than 31.4 C, or 88.5 F from days with lower maximum temperatures). Of those bins characterized by cooler temperatures, only Bin 7 is an exceedance bin. Bin 7 contains 9 exceedance days (it also contains 7 non-exceedance days). Days within this bin are characterized within the CART analysis by a maximum surface temperature between 29.1 and 31.4 C (84.3 and 88.5 F), low relative humidity at midday (less than 53.5 percent), low to moderate wind speeds at 850 mb (the daily average value is less than 3.99 ms-1 but greater than 1.82 ms-1), and low surface wind speeds. The maximum observed ozone concentration at any site is also more than 17.6 percent higher than the average maximum over all sites. This indicates that the observed ozone peak is relatively isolated, and it is likely that an exceedance was observed at only one site. Days within this bin are distinguished from those in other bins by surface wind speeds and previous days ozone maxima. Higher wind speeds during the morning hours, for example, result in designation of Bin 8 as belonging to Category 3 (maximum ozone between 105 and 125 ppb). Higher ozone on the previous day distinguishes days within Bin 7 from those in Bin 6.

Compared to other exceedance bins, the average 850 mb wind speeds are generally lower for Bin 7 than the other bins (an exception is Bin 19). As noted earlier, the average maximum temperature is lower for this bin than the other exceedance bins.

A review of all of the CART input variables for days classified within this bin confirms that these days are characterized by low winds speeds near the surface (throughout the day), low wind speeds at the 850 mb level, and moderate wind speeds at the 700 mb level, low relative humidity, and very little cloud cover. All of these indicate favorable conditions for local ozone production and little potential (due to the light wind speeds) for significant ozone or precursor transport. The timing of the maximum ozone concentration (1200 and 1600 LST) also suggests a dominant home-grown component. The maximum observed ozone concentration is on the average 25 percent higher than the average maximum concentration. This indicates a relatively isolated ozone peak, again suggesting a predominantly local (and thus spatially variable) contribution.

The true exceedance days within this bin are characterized by less cloud cover, especially during the morning hours, than other days within the bin. These days also are distinguished by a higher maximum ozone concentration on the previous day, suggesting some build-up and possible recirculation of ozone within the area.

Within the ten-year period of record, there are three occurrences of the bin sequence 3-7. Days within Bin 3 are characterized by moderate ozone concentrations. This supports the possibility of recirculation.

These results indicate that exceedance days within this bin are primarily the result of local emissions and ozone conducive meteorological conditions. Transport likely did not contribute significantly to the observed ozone exceedances.

Bin 13

One exceedance day is incorrectly placed in Bin 13 (which is a Category 3 terminal bin). Note that the maximum observed ozone concentration for this day is 132 ppb. Days within this bin are characterized by a maximum surface temperature greater than 31.4 C (88.5 F), low relative humidity at midday (less than 41.5 percent), low to moderate surface wind speeds, and a maximum ozone concentration of less than approximately 89 ppb on the previous day. This is a neighbor to Bin 15 (an exceedance bin that is described below). These two bins are distinguished, according to CART, by the previous days maximum ozone concentration (that for Bin 13 is lower). Considering the remaining input variables, nearly all indicate less ozone conducive meteorological conditions for Bin 13. The average wind speed at the 700 mb level is more than 50 percent higher for the Bin 13 days, suggesting the potential for pollutant transport aloft. The ratio of the maximum to the average maximum ozone concentration is relatively low, indicating widespread ozone concentrations, which is also indicative of transport.

These characteristics suggest that high ozone concentrations for days within this bin are the combined result of local ozone production and pollutant transport aloft. One of the days placed in this bin is 7 July 1988, which has been studied extensively and is considered to be a day on which ozone transport from the Midwest and the northeastern U.S. into the Atlanta area occurred (Douglas et al., 1994).

Bin 15

Bin 15 contains 20 exceedance days (it also contains 12 non-exceedance days; half of these represent days from Category 3). Days within this bin are characterized within the CART analysis by a maximum surface temperature greater than 31.4 C (88.5 F), low relative humidity at midday (less than 41.5 percent), low to moderate surface wind speeds during the afternoon hours (the daily average value for the hours 1300-1600 LST is less than 4 ms-1). Wind speeds aloft are also low, especially at the 850 mb level. The previous days maximum observed ozone concentration is between approximately 89 and 118 ppb. Days within this bin are distinguished from those in other bins by this last parameter. Days with lower ozone concentrations during the previous day were classified by CART as belonging to Category 3 (maximum ozone between 105 and 125 ppb).

Compared to the other exceedance bins, Bin 15 is characterized by low relative humidity, high temperatures (the average is 34.4 C or 94 F), and relatively widespread high ozone concentrations (the percentage difference between the maximum and the average maximum ozone concentration is relatively small compared to the other exceedance bins).

A review of all of the CART input variables for days classified within this bin suggests that the low wind speeds near the surface and aloft, high temperature, and low relative humidity resulted in favorable conditions for local ozone production. Some potential for carry-over or recirculation is indicated by the requirement for the previous days maximum ozone concentration to be greater than approximately 89 ppb. Day-to-day carryover may also explain the widespread high ozone conditions.

The true exceedance days within this bin are characterized by slightly more stable conditions throughout the lower atmosphere than other days within the bin. Otherwise, the average conditions are very similar.

The data indicate that exceedance days within this bin are the combined result of local ozone production under ozone conducive meteorological conditions and day-to-day carryover of ozone. Note, however, that this carryover could include a transport component. Specifically, two exceedance days within Bin 15 are preceded by days within Bin 13, which appear to have some contribution from transport. One of the days placed in this bin is 8 July 1988, which has been studied and modeled extensively. This description is consistent with previous analysis of this episode (Douglas et al., 1994).

Bin 16

One exceedance day is misclassified as belonging to Category 2 and is listed as a member of Bin 16. Key meteorological characteristics of days within this bin (according to the CART splits) include somewhat higher relative humidity compared to Bins 13 and 15 and a lower percentage difference between the maximum and average maximum ozone concentration. Lower temperatures and cloud cover also contribute to the less ozone conducive meteorological conditions. However, surface wind speeds are relatively low and the previous days maximum ozone concentration is 85 ppb, when averaged over all days within the bin. So some potential for ozone formation and carryover exist. Bin 16 contains twelve pairs for which a Bin 16 day is followed by a Bin 16 day, this repeating pattern also suggests carryover. Based on these characteristics, the exceedance day is classified as predominantly home grown. Low to moderate wind speeds aloft indicate a slight potential for contribution from transport.

Bin 17

Two exceedance days were classified within Bin 17, another Category 3 terminal bin and a neighbor to Bin 16. Days within this bin are characterized by a maximum surface temperature greater than 31.4 C, relative humidity at midday that is greater than 41.5 percent, low wind speeds near the surface and aloft (with limits of 2.36 ms-1 for the surface between 0700 and 1000 LST, 4 ms-1 for the surface between 1300 and 1600 LST, and 2.96 ms-1 at the 850 mb level). The previous days maximum ozone concentration is below 118 ppb and the percentage difference between the peak and the average peak is greater than 12 percent.

Days grouped within Bin 17 exhibit some of the lowest wind speeds compared to those within the other bins, indicating that these are true stagnation events. Higher relative humidity than other Category 3 or 4 bins is associated with higher amounts of cloud cover and this likely leads to the lower observed ozone concentrations. Note that the maximum ozone concentration reached 128 ppb and 132 ppb, on the two exceedance days classified in this bin.

Due the stagnation conditions, the exceedance days within this category are most likely home-grown events.

Bin 18

Bin 18 contains 12 exceedance days (it also contains 14 non-exceedance days, half of which represent the next lower category of ozone concentrations). According to CART, the distinguishing characteristics of this bin are the same as those for bin 17, with one exception. The average wind speed at the 850 mb level is greater than 2.96 ms-1, higher than that for Bin 17. A comparison of the characteristics of the two bins, however, shows that wind speeds are indeed generally higher for Bin 18, especially at the surface during the afternoon (1300-1600 LST) and aloft. Maximum temperature, relative humidity, and cloud cover are lower for Bin 18. Thus, stagnation is not as prevalent.

Compared to the other exceedance bins, surface wind speeds are lower but wind speeds aloft are slightly higher. The percentage difference between the maximum and average maximum ozone concentration is relatively high, suggesting an isolated peak. This is consistent with the low surface wind speeds. Less cloud cover distinguishes exceedance days within this bin from non-exceedance days within this bin, suggesting local ozone formation.

The analysis results indicate that exceedance days within this bin are associated with ozone conducive meteorological conditions. Accordingly, these days are classified as predominantly home grown. However, the moderate wind speeds aloft coupled with the relatively low (compared to other exceedance bins) maximum ozone concentration on the previous day indicate some potential for a contribution from transport.

Bin 19

Bin 19 contains 4 exceedance days (it also contains 2 non-exceedance days). According to CART, the distinguishing characteristics of this bin are similar to those for bins 17 and 18, with the exception that the average morning surface wind speed (0700-1000 LST) is greater than 2.36 ms-1 and the average wind speed at the 850 mb level is very low (less than 1.75 ms-1). The maximum ozone concentration on the previous day is also lower for this bin. Days within this bin are also characterized by high temperatures (the average maximum temperature for the bin is 33.3 C or 92 F).

Compared to other exceedance bins, surface wind speeds are higher during the early morning hours, lower during the late morning hours, and similar during the afternoon hours. Wind speeds aloft, however, are notably lower for this bin. Consequently, exceedance days within this bin are classified as predominantly home grown.

The data suggest that, while some recirculation is possible, high ozone concentrations for days within this bin are primarily the result of stagnation conditions near the surface and aloft. High temperatures and clear skies support the conclusion regarding ozone conducive meteorological conditions.

Bin 21

Higher wind speeds aloft (850 mb level) distinguish days within this bin from those within Bin 19. This bin is designated a Category 3 bin but contains one exceedance day. The characteristics of days within this bin are similar to those for Bin 19 and a predominant home grown component is indicated. Higher ozone concentrations on the previous day suggest that recirculation is more likely. The higher wind speeds aloft coupled with greater amounts of cloud cover suggest that this exceedance day has a slightly greater potential contribution from transport (and less conducive meteorological conditions) than those placed within Bin 19.

Bin 26

Two exceedance days are classified as belonging to Bin 26, another Category 3 bin. Days within this bin are characterized by relatively high wind speeds (especially compared to the exceedance bins) both near the surface and aloft. High ozone concentrations (in excess of 102 ppb) were measured in the Atlanta area on each day prior to those within the bin. High temperature was also recorded within the area (the average maximum temperature for all days within this bin is greater than approximately 34 C (93 F)).

While certain ozone conducive meteorological conditions occurred, the relatively high wind speeds for this bin suggest that pollutant transport was possible throughout the day both near the surface and aloft. Thus, exceedance days within this bin are classified as potential transport days.

Bin 28

Bin 28 contains 51 exceedance days (it also contains 19 non-exceedance days). According to CART, the distinguishing characteristics of this bin include a maximum temperature greater than 31.4 C, maximum ozone on the previous day greater than 118 ppb, and relatively low surface wind speeds during the afternoon hours (1300-1600 LST) as well as at the 850 mb level. A higher peak to average maximum ozone ratio distinguishes this bin from neighboring bins.

Compared to the other exceedance bins, wind speeds aloft (at both the 850 and 700 mb levels) are typically lower (an exception to this is Bin 19 which was characterized by stagnation). The maximum temperature is highest of any of the exceedance bins; the average value is approximately 35 C (95 F). The previous days maximum ozone concentration is also high compared to most of the other bins and is on the average approximately 144 ppb. Combined with widespread high ozone concentration, these factors strongly suggest that days within this bin are home-grown episodes with some potential for day-to-day carryover or recirculation. Surface wind speeds are low (low to moderate during the period 1300-1600 LST). Low relative humidity and light cloud cover round out the conditions associated with this exceedance bin.

The analysis results indicate that exceedance days within this bin, which comprise nearly half of the observed exceedance days that occurred during the ten-year analysis period, are associated with local ozone conducive meteorological conditions and occur during a multi-day ozone episode. High ozone concentrations on the previous day appear to be an important characteristic of days within this bin. Thus the potential for day-to-day build up or carryover of ozone is implied. There are thirty pairs of consecutive days (two-day periods) in Bin 28 and several other longer sequences of days with this pattern. Thus, conditions associated with this bin are not only prevalent throughout the analysis period but are persistent as well. Two days within this bin follow a day within Bin 26, which is designated a moderate transport bin. The remaining days appear to be primarily influenced by local recirculation. Days within this bin are therefore classified as home-grown (and possibly influenced by day-to-day carryover of ozone within the Atlanta area).

Three days from the 1987 State Implementation Plan (SIP) modeling episode are placed in this bin: 30 July, 31 July, and 1 August. Classification as a home-grow episode with recirculation is consistent with the more detailed analysis of the episode provided by Douglas et al. (1994).

Bin 30

Bin 30 contains 2 exceedance days and 1 non-exceedance day. CART separates these days from those within Bin 28 based on higher wind speeds aloft (at the 850 mb level). The classification tree further segregates the days within this bin from other days with the higher wind speeds aloft based on very low relative humidity.

Compared to the other exceedance bins, wind speeds near the surface and aloft are higher. This is especially true aloft where the average wind speed is greater than 6 ms-1 for both the 850 and 700 mb levels. These are the highest average wind speeds aloft for any of the exceedance bins. Interestingly, the previous days maximum ozone concentration is also the highest among any of the exceedance bins, relative humidity is the lowest, and the lower atmosphere is most stable.

The true exceedance days within this bin are characterized by slightly lower wind speeds and less stable conditions. The percentage difference between the peak and average maximum ozone concentration is higher for the exceedance days. Recall, however, there are only three days within the bin.

The analysis results indicate that although ozone conducive meteorological conditions exist, the potential for a contribution from transport is higher than for days within bin 28, due to the higher wind speeds aloft. The two exceedance days within this bin, therefore, appear to be the combined result of local ozone production and transport. According to the designations established at the beginning of this section, exceedance days within this bin are classified as potential transport days.

Bin 32

Higher relative humidity and greater amounts of cloud cover distinguish days within Bin 32 from those within Bin 30. This bin is designated a Category 3 bin but contains one exceedance day. The overall characterization of days within this bin relative to the potential contribution from transport is, however, similar. Local meteorological conditions are less conducive to ozone production and overall ozone concentrations are lower. This comparison, however, confirms the relative importance of the home-grown component of exceedance days within Bin 30.

Summary of Classification of Exceedance Days for Atlanta

The results of the CART-based episode classification analysis for the Atlanta area are summarized in Table 3-2. This table presents information about both exceedance bins and non-exceedance bins that contain misclassified exceedance days. For each bin, the importance of day-to-day carryover or recirculation of ozone was estimated primarily based on the previous days ozone concentration and wind speed aloft. Determination of low, moderate, and high wind speeds was based on the distribution of observed values throughout the study period. However, the other characteristics of days within the bin, as described in the preceding text, were also considered. The potential for contribution from transport was estimated primarily based on surface and upper-air wind speeds. Again, other factors as described above were used to confirm the characterization.

For most of the exceedance days considered in this analysis, day-to-day carryover or recirculation is moderately important. This typically corresponds to moderate ozone concentrations on the previous day and low wind speeds or high ozone concentrations on the previous day and moderate wind speeds. For Bin 13, high wind speeds aloft and low ozone on the previous day render recirculation less likely. For Bin 28, low to moderate wind speeds aloft coupled with high ozone concentrations on the previous day enhance the potential for recirculation.

The potential for contribution from transport ranges from low to moderate for the identified exceedance days (corresponding to low to moderate wind speeds both near the surface and aloft). This potential is enhanced for Bins 13 and 30 by high wind speeds aloft and for Bins 26 and 32 by high wind speeds both aloft and near the surface. In most cases, the higher wind speeds are also coupled with generally less ozone conducive meteorological conditions (such as lower temperatures and higher relative humidity). Days with a moderate or greater transport potential are identified in Table 3-2 as transport days. All other days are classified as predominantly home grown.

In summary, of the 107 exceedance days considered in this analysis 101 days are predominantly home grown and 6 are characterized by a moderate potential for a contribution from transport. The potential transport days occurred at a average rate of less than one per year during the ten-year period. Considering all possible three-year periods within the analysis period, a maximum of three such days occurred within any three-year period.

 

Table 3-1. Summary of classification accuracy for the Atlanta CART Tree.

True Class

 

 

1

2

3

4

C

1

1116

55

0

0

A

2

300

221

38

1

R

3

44

66

90

8

T

4

9

24

22

98

 

 

TABLE 3-2. Summary of classification of exceedance days for Atlanta.

 

 

Bin No.

 

Exceedance Bin

No. of Exceedance Days

Average Maximum Ozone (ppb)

Importance of Recirculation or Day-to-Day Carryover

Potential for Contribution from Transport

Classified as Predominantly "Home Grown"

2

N

1

130.0

MH

L

Y

7

Y

9

134.7

MH

L

Y

13

N

1

132.0

LM

M

N

15

Y

20

142.8

MH

L

Y

16

N

1

135.0

M

LM

Y

17

N

2

130.0

MH

L

Y

18

Y

12

131.0

M

LM

Y

19

Y

4

129.8

M

L

Y

21

N

1

135.0

MH

L-M

Y

26

N

2

145.7

M

M

N

28

Y

51

148.3

H

L

Y

30

Y

2

134.0

MH

M

N

32

N

1

126.0

M

M

N

Key: Y=Yes, N=No, H=High, M=Moderate, L=Low

 

 

Predominantly "Home Grown"

 

Influenced by Transport

No. of Days

101

 

6

Average Maximum Ozone (ppb)

135.2

 

134.4

BIRMINGHAM

The results of the CART analysis for the Birmingham area are discussed in this section. This discussion is followed by a detailed analysis of the meteorological and transport characteristics of the days within the exceedance bins

As indicated in the previous section, key classification parameters include wind speeds (near the surface and aloft) and the previous days maximum ozone concentration. In the following summary for the Birmingham area, the descriptors low, moderate, and high for surface wind speeds refer to those that are less than 2.5 ms-1, between 2.5 and 4 ms-1, and greater than 4 ms-1, respectively. For the upper levels, they refer to wind speeds that are less than 4 ms-1, between 4 and 9 ms-1, and greater than 9 ms-1, respectively. For ozone, the descriptors refer to concentrations that are less than 85 ppb, between 85 and 105 ppb, and greater than 125 ppb, respectively.

CART Results

The classification tree for Birmingham consists of 33 terminal bins (Figure 3-2). Of these, 4 bins are designated exceedance bins that contain all days on which the maximum ozone concentration at any regulatory site within the Birmingham urban area exceeded the current federal ozone standard of 124 ppb. The classification accuracy of the CART tree is depicted in Table 3-3. Good classification is achieved within most categories. During the ten-year period included in this analysis, there are only 12 exceedance days. All of these are properly classified as having maximum ozone concentrations that are greater than or equal to 125 ppb (Category 4). There are also 12 non-exceedance days misclassified into the exceedance category.

Certain variables were identified by the CART program as being of primary importance in constructing the binary classification tree for the Birmingham area. These include yesterdays maximum ozone concentration, average surface wind speed between 1000 and 1300 LST, average wind speed at the 850 mb level, relative humidity at noon, average surface wind speed between 0700 and 1000 LST, the percentage difference between the maximum ozone concentration at any of the regulatory sites and the average maximum over all of these sites, average surface wind speed between 1300 and 1600 LST, and the daily maximum surface temperature.

Review of these important parameters and the classification tree (the parameter splits and the average values of the input parameters associated with each terminal bin) indicates that the relationship between the meteorological variables and the ozone concentrations depicted by the CART tree is physically meaningful. For example, considering all bins within each of the four classification categories 1 through 4, corresponding to increasing maximum ozone concentrations, the average daily maximum temperature increases with each higher category. Conversely, relative humidity and surface wind speed generally decrease with each higher category.

Exceedance Bin Characteristics

Bin 10

Bin 10 contains one exceedance day and no non-exceedance days. According to CART, the distinguishing characteristics of this bin are similar to those for an adjacent Category 3 bin (Bin 11), with the exception that the average morning surface wind speed (0700-1000 LST) is very low (less than or equal to 0.28 ms-1). The day within this bin is also characterized by a maximum surface temperature between 30.3 and 37.5 C (86.5 and 99.5 F), low relative humidity at noon (less than or equal to 44.5 percent), and low surface wind speeds between 1000 and 1300 LST (less than or equal to 1.68 ms-1). The exceedance day is also characterized by a peak ozone concentration that is approximately 23 percent higher than the average maximum concentration.

Compared to other exceedance bins, surface wind speeds are significantly lower during all periods and the difference between the maximum ozone concentration is higher than the average maximum concentration.

The data suggest that the high ozone concentration for the day is primarily the result of stagnation conditions near the surface. The relatively large difference between the maximum and average maximum concentration supports this conclusion. High temperatures and predominantly clear skies reflect the ozone conducive meteorological conditions. The potential for recirculation, however, is also indicated.

Bin 15

Bin 15 also contains only one exceedance day and no non-exceedance days. According to CART, the distinguishing characteristics of this bin are similar to those for an adjacent Category 3 bin (Bin 14), with the exception that the difference between the temperature aloft and at the surface during the morning hours is greater, thereby signifying more stable conditions. As for Bin 10, the day within this bin is characterized by a maximum surface temperature between 30.3 and 37.5 C (86.5 and 99.5 F) and low relative humidity at midday (less than or equal to 44.5 percent). Wind speeds at the surface between 0700 and 1000 LST are low but somewhat higher than for Bin 10 (1.17 to 2.73 ms-1) as are those between 1000 LST and 1300 LST (1.68 to 3 ms-1). This bin is also characterized by average wind speeds for the afternoon period of 1300 to 1600 LST of 3.43 ms-1 and an average wind speed aloft of 3.5 ms-1. The difference between the temperature aloft and at the surface for this bin is greater than -1.2 C. This last meteorological feature indicates relatively stable conditions within the lower atmosphere and distinguishes this bin from the adjacent Category 3 bin (Bin 14).

Compared to other exceedance bins, surface wind speeds for the period of 1300 to 1600 LST are slightly higher as are the wind speeds at 700 mb. Sky conditions are also clear for the three time periods considered.

The clear skies, low wind speeds, and stability result in ozone conducive meteorological conditions and the high ozone concentrations. While the potential influence from transported ozone is greater than that for Bin 10, due to the low wind speeds, it does not appear to be a significant contributor.

Bin 16

Bin 16 contains nine exceedance days and twelve non-exceedance days. Based on the criteria selected by CART, this bin differs from the adjacent Category 4 bin (Bin 15) in that the average wind speed at 850 mb is greater than 1.95 ms-1 and since this condition requires a different path to be followed, no restriction is placed on the temperature lapse rate. An analysis of the average temperature lapse rate indicates somewhat more instability and hence a possible explanation for the misclassification of the non-exceedance days. Although CART did not specify clouds as being important, examination of the cloud data indicates that misclassification may also have been due to cloudy conditions. Clouds were present on both the exceedance and non-exceedance days. The meteorological feature that distinguishes this bin from the adjacent Category 3 bin (Bin 17) is that the average wind speed at the 850 mb level is greater than 6.58 ms-1 for Bin 17. Winds aloft for Bin 16 (this bin) are lower; the average values are 3.46 ms-1 at the 850 mb level and 4.4 ms-1 at the 700 mb level. Surface wind speeds are low during the period 0700-1300 LST and slightly higher during the afternoon period (4.4 ms-1).

Compared to other exceedance bins, sky conditions tend to be cloudier. Nevertheless, the maximum temperature is relative high and the time of the maximum ozone concentration is slightly earlier (suggesting the possibility of day-to-day carryover).

Although cloudy conditions complicate the analysis, the low wind speeds and high temperatures support the classification of home grown. Some day-to-day recirculation is indicated. A slight potential for a contribution from transport is also suggested.

Bin 23

Bin 23 contains one exceedance day and no non-exceedance days. According to CART, the distinguishing characteristics of this bin are similar to those for an adjacent Category 3 bin (Bin 2), with the exception that the height of the 700 mb level is greater than 3220 m. In addition to the criteria for the height of the 700 mb level, the day within this bin is characterized by a maximum surface temperature of greater than 30.3 C (86.5 F), very low wind speeds at the surface between 0700 and 1000 LST (less than 0.59 ms-1), low wind speeds between 1000 and 1300 LST (less than 3.0 ms-1), and a relative humidity at midday or greater than 44.5 percent.

Compared to the other exceedance bins, the average temperature lapse rate is the greatest (-5.8 C), average maximum temperature the highest, and maximum ozone concentration on the preceding day, the highest. In addition, clouds were indicated for the 1400 LST period.

The data suggest that high ozone concentrations for days within this bin are primarily the result of stagnation conditions near the surface. High temperature and pressure support the conclusion regarding ozone conducive meteorological conditions. Upper-air wind data were missing and therefore no conclusion could be made regarding transport except that, based on surface wind speed alone, the potential for transport could be no greater than moderate.

Summary of Classification of Exceedance Days for Birmingham

The results of the CART-based episode classification analysis for the Birmingham area are summarized in Table 3-4. For each bin, the importance of day-to-day carryover or recirculation of ozone was estimated primarily based on the previous days ozone concentration and wind speed aloft. Determination of low, moderate, and high wind speeds was based on the distribution values throughout the study period.3 However, the other characteristics of days within the bin, as described in the preceding text, were also considered. The potential for contribution from transport was estimated primarily based on surface and upper-air wind speeds.4 Again, other factors as described above were used to confirm the characterization.

For most of the exceedance days considered in this analysis, day-to-day carryover or recirculation is of moderate importance. For Bin 10, the ozone concentration on the previous day is moderate (between 85 and 125 ppb) and the average wind speed is low. These conditions combine to produce a moderate to high potential for recirculation. For Bins 16 and 15, low ozone concentrations on the previous day coupled with moderate and low wind speeds aloft, combine to produce low-to-moderate and moderate potentials for recirculation, respectively. Classification for Bin 23 using this method could not be made due to the missing wind data aloft (only one day fell into this bin). Based on the ozone concentration on the previous day, classification could be no greater than moderate-high and this could only be the case if the wind speed aloft exceeded 4 ms-1.

For most exceedance days, the potential for contribution from transport ranges from low to moderate (corresponding to low to moderate wind speeds both near the surface and aloft). Average wind speeds aloft and at the surface for Bin 16 are approximately equal to (4.13 and 2.503 ms-1, respectively) the cutoff values for moderate. Considering the other parameters, a classification of low-to -moderate was applied. Because of missing wind speed data aloft for the day in Bin 23, this bin could not be classified, however, based on the surface wind speed, classification could be no greater than moderate.

None of the 12 exceedance days studied in this analysis can be characterized as having a dominant pollutant transport component.

 

Table 3-3. Summary of classification accuracy for the Birmingham CART Tree.

True Class

 

 

1

2

3

4

C

1

1427

23

1

0

A

2

350

123

6

0

R

3

83

39

64

0

T

4

7

5

0

12

 

 

 

TABLE 3-4. Summary of classification of exceedance days for Birmingham.

 

 

Bin No.

 

Exceedance Bin

No. of Exceedance Days

Average Maximum Ozone (ppb)

Importance of Recirculation or Day-to-Day Carryover

Potential for Contribution from Transport

Classified as Predominantly "Home Grown"

10

Y

1

146.0

MH

L

Y

15

Y

1

138.0

LM

LM

Y

16

Y

9

133.3

MH

LM

Y

23

Y

1

139.0

no greater than

MH

no greater than

M

--

Key: Y=Yes, N=No, H=High, M=Moderate, L=Low

 

 

Predominantly "Home Grown"

 

Influenced by Transport

No. of Days

12

 

0

Average Maximum Ozone (ppb)

139.1

 

--

 

CHARLOTTE

The results of the CART analysis for the Charlotte area are discussed in this section. This discussion is followed by a detailed analysis of the meteorological and transport characteristics of the days within the exceedance bins. For the Charlotte area, only one day on which an exceedance of the federal ozone standard was recorded was not classified correctly into an exceedance bin; this additional (Category 3) bin is also examined.

As indicated in the previous section, key classification parameters include wind speeds (near the surface and aloft) and the previous days maximum ozone concentration. In the following summary for the Atlanta area, the descriptors low, moderate, and high for surface wind speeds refer to those that are less than 2.5 ms-1, between 2.5 and 4 ms-1, and greater than 4 ms-1, respectively. For the upper levels, they refer to wind speeds that are less than 4.5 ms-1, between 4.5 and 10 ms-1, and greater than 10 ms-1, respectively. For ozone, the descriptors refer to concentrations that are less than 85 ppb, between 85 and 105 ppb, and greater than 125 ppb, respectively.

CART Results

The classification tree for Charlotte consists of 33 terminal bins (Figure 3-3). Of these, 7 bins are designated exceedance bins and contain mostly days for which the maximum ozone concentration at any site within the Charlotte area exceeded the current federal ozone standard of 124 ppb. The classification accuracy of the CART tree is depicted in Table 3-5. As for the Atlanta and Birmingham areas, good classification is achieved within most categories and, in most cases, days that are misclassified are put into an adjacent category. During the ten-year period included in this analysis, there are 36 exceedance days. Thirty five of these are properly classified as having maximum ozone concentrations that are greater than or equal to 125 ppb (Category 4). One exceedance day is misclassified as having a maximum ozone concentration greater than or equal to 105 ppb but less than 125 ppb (Category 3). Twenty nine days are misclassified as exceedances indicating that the conditions associated with the ozone exceedance days are incompletely described by the CART parameters. Ten of these days are misclassified into Category 3 and are characterized by relatively high, although not exceedance level, ozone concentrations. Of the remaining days, more than half are weekend days or Mondays, indicating that the misclassification may be due to lower weekend emissions.

Certain variables were identified by the CART program as being of primary importance in constructing the binary classification tree for the Charlotte area. These include yesterdays maximum ozone concentration, relative humidity at noon, average surface wind speed between 1000 and 1300 LST, average height of the 700 mb surface, average surface wind speed between 0700 and 1000 LST, maximum temperature, average wind speed at the 850 mb level, average surface wind speed between 1300 and 1600 LST, and cloud cover at 1000 LST.

Review of these important parameters and the classification tree (the parameter splits and the average values of the input parameters associated with each terminal bin) indicates that the relationship between the meteorological variables and the ozone concentrations depicted by the CART tree are physically meaningful but different from that for the Atlanta and Birmingham areas. For example, considering all bins within each of the four classification categories 1 through 4, corresponding to increasing maximum ozone concentrations, the average daily maximum temperature increases from Category 1 to Category 3 but decreases slightly for Category 4. Relative humidity and cloud cover decrease with each higher category. Variations in wind speed at the surface and aloft vary among the categories in a manner that is consistent with maximum temperature. The implications of this with respect to the contribution from transport are investigated below.

Exceedance Bin Characteristics

Bin 8

The first eleven terminal bins of the Charlotte CART tree contain those days with relatively low daily maximum surface temperatures (the first branching separates those days with a maximum temperature greater than 32.5 C or 90.5 F from days with lower maximum temperatures). Of those bins characterized by cooler temperatures, only Bin 8 is an exceedance bin. Bin 8 contains 3 exceedance days and no non-exceedance days. Days within this bin characterized by a maximum surface temperature between 30.8 and 32.5 C (87.4 and 90.5 F), low surface winds speeds during the (less than 2.04 ms-1 during 1000-1300 LST), and moderate wind speeds aloft. Despite the relatively low temperatures and the moderate to high wind speeds aloft, the average maximum ozone concentration for this bin (averaged over all sites included in the analysis and all days within the bin) is among the highest of the exceedance bins (138 ppb). Days within this bin are distinguished from those in a neighboring bin by the percentage difference between the maximum ozone concentration at any site and the average maximum concentration. Days with a substantially lower percentage difference belong to Category 2 (Bin 7)

Compared to other exceedance bins, the average maximum temperature is lower for Bin 8 than the other bins, the average 850 mb wind speeds are generally higher, and the morning sounding indicates less stable conditions.

A review of all of the CART input variables for days classified within this bin confirms that these days are characterized by low winds speeds near the surface (throughout the day), moderate wind speeds at the 850 and 700 mb levels, and clear skies. The average maximum ozone concentration on the day prior to each day within this bin is approximately 92 ppb. That this bin is distinguished from a Category 2 bin by the percentage difference between the maximum and average ozone concentration, suggests that local stagnation played a dominant role in the exceedances. Other factors indicate that carryover or recirculation also occurred. Overall, days within this bin are classified as predominantly home grown, however, some potential for pollutant transport is indicated.

Bin 13

Bin 13 contains 4 exceedance days and no non-exceedance days. Days within this bin are characterized, according to CART, by a maximum surface temperature greater than 34.7 C (94.5 F), very low surface winds speeds especially during the afternoon hours, and relatively unstable lapse rates during the morning hours. The high temperatures and less stable conditions distinguish days within this bin from days within nearby (Category 3) bins.

The average maximum temperature for Bin 13 is the highest among the exceedance bins. Wind speeds near the surface and aloft are also relatively low compared to most other exceedance bins.

A review of all of the CART input variables for days classified within this bin confirms that these days are characterized by high temperatures (the average maximum temperature for the bin is 35.9 C (96.6 F), very low winds speeds near the surface (throughout the day, but especially during the afternoon hours), low wind speeds at the 850 mb level, low to moderate wind speeds at the 700 mb level, and relatively unstable conditions within the lower atmosphere during the morning hours. The average maximum ozone concentration on the day prior to each day within this bin is 98 ppb. Some cloud cover was observed at 1400 LST.

The characteristic conditions associated with days within this bin strongly indicate local stagnation with the potential for some day-to-day carryover or recirculation. Transport is not indicated.

Bin 15 

One exceedance day is incorrectly placed in Bin 15 (which is a Category 3 terminal bin). Days within this bin are characterized by maximum temperatures greater than 32.5 C (90.5 F), relatively unstable conditions within the lower atmosphere during the morning hours, and very light wind speeds during the period 1000-1300 LST. These characteristics distinguish days within this bin in the CART analysis. A review of all of the CART input parameters, indicates that low wind speeds prevail near the surface throughout the day and the wind speeds aloft are low to moderate. High ozone concentrations on the previous day are also characteristic of days within this bin. While the average maximum temperature is comparable, cloud cover is greater for this bin than any of the exceedance bins (this is likely the cause of the overall lower ozone concentrations compared to the exceedance bins).

Clearly the low surface wind speeds and high temperatures contribute to local ozone production within the Charlotte area. High ozone on the previous day and the relatively unstable morning vertical temperature gradient suggests the potential for day-to-day carryover. Days within this bin are classified as predominantly home grown. Some moderate wind speeds aloft, however, also provide a mechanism for transport.

Bin 19

Bin 19 contains 4 exceedance days (it also contains 8 non-exceedance days). CART distinguishes days within this bin by relatively stable conditions during the morning hours, lower ozone concentrations on the previous day compared to nearby exceedance bins, and low relative humidity at midday (less than 38.5 percent). Other characteristic of days within this bin (considering all variables) include clear skies, low wind speeds near the surface, and moderate wind speeds aloft.

Compared to most of the other exceedance bin for the Charlotte area, surface winds speeds for this bin are slightly higher (with the exception of Bin 31) while relative humidity and cloud cover are lower.

The true exceedance days within this bin exhibit lower surface wind speeds but higher upper-air wind speeds (on average) than other days in the bin. Thus, while most indicators suggest that local conditions are conducive to ozone formation, some contribution from transport aloft (as well as near the surface) is also indicated. As a result, the classification of days within this bin reflects moderate potential for a contribution from pollutant transport.

Bin 20

Bin 20 contains 2 exceedance days (it also contains 4 non-exceedance days). CART segregates days within this bin based on a number of factors including previous days' maximum ozone concentration, morning temperature gradient, relative humidity, and surface wind speed. As for Bin 19, stable conditions during the morning hours and relatively low ozone on the previous day characterize days within this bin. However, continuing with this comparison, relative humidity is higher and cloud cover is greater for days placed in Bin 20 while surface wind speeds are lower. Days within both bins are characterized by moderate wind speeds aloft.

High relative humidity and cloud cover distinguish days within this bin from those within all other exceedance bins. The average percentage difference between the maximum and average maximum ozone concentration is higher than all other bins; this feature is indicative of local ozone formation.

Surface level stagnation appears to be the primary contributor to the ozone exceedances. As for Bin 19, wind speeds aloft are higher for the true exceedance days within the bin. Thus, some potential for transport is also indicated.

Bin 27

Bin 27 contains 8 exceedance days (it also contains 8 non-exceedance days). In addition to maximum temperature, the CART tree separates days within this bin according to stability (days within this bin are relatively stable), low relative humidity (less than 54.5 percent), sparse cloud cover, and low wind speeds aloft. Other characteristics of days within this bin include low surface wind speeds and high ozone on the previous day.

Compared to the other exceedance bins for the Charlotte area, winds speeds aloft are the lowest for Bin 27. The previous days' maximum ozone concentration is also among the highest (only that for Bin 31 is higher). The maximum ozone concentration also occurs earlier (on the average) for days within this bin.

These conditions strongly suggest that days within this bin are part of a multi-day ozone episode during which day-to-day carryover exists. In particular high ozone on the previous day, the light winds aloft, and the early timing of the peak concentration (a signature of carryover) support this conclusion. Transport is not indicated. Days within this bin are classified as home grown.

Bin 29

Bin 29 contains 13 exceedance days (it also contains 9 non-exceedance days). The characteristics of this bin are similar to those for Bin 27, except that wind speeds aloft are higher. CART identifies wind speed at the 700 mb level as an important distinguishing feature of the bin (and one that distinguished data within this bin from days within Bin 27). The peak ozone concentration also occurs later in the day.

The moderate wind speeds aloft associated with this bin (average values are 5.4 ms-1 for the 850 mb level and 8 ms-1 for the 700 mb level) suggest that day-to-day carryover is less of an important contributor to the ozone exceedances than for Bin 27 but that transport is more important. As for Bin 27, local conditions are conducive to ozone production.

Bin 31

Bin 31 contains one exceedance day and no non-exceedance days. This day is characterized by a strong morning inversion (stable conditions), low to moderate wind speeds near the surface and aloft, somewhat higher relative humidity than the other exceedance days, but no cloud cover. The previous day's ozone concentration is very high (169 ppb compared to 127 ppb for the exceedance day within this bin). CART identifies yesterday's maximum ozone concentration as the important parameter that distinguished this day from those in Bin 30 (which is a Category 3 bin). This suggests that the exceedance day in this bin is a clean out day for a severe ozone episode and is, therefore, attributable to local emissions and day-to-day carryover. Surface wind speeds that increase throughout the day and a lower maximum temperature than most other exceedance days support this conclusion.

Summary of Classification of Exceedance Days for Charlotte

The results of the CART-based episode classification analysis for the Charlotte area is summarized in Table 3-3. This table presents information about the exceedance bins, as well as the Category 3 bin that contained an exceedance day. For each bin, the importance of day-to-day carryover and the potential for contribution from transport were estimated. The importance of day-to-day carryover or recirculation of ozone was estimated primarily based on the previous days ozone concentration and wind speed aloft. Determination of low, moderate, and high wind speeds was based on the distribution values throughout the study period. The other characteristics of days within the bin, as described in the preceding text, were also considered. The potential for contribution from transport was estimated primarily based on surface and upper-air wind speeds. Again, other factors as described above were used to confirm the characterization.

For most of the exceedance days considered in this analysis, day-to-day carryover or recirculation is moderately important. This typically corresponds to moderate to high ozone concentrations on the previous day and low to moderate wind speeds aloft. For Bins 19 and 20, relatively low maximum ozone concentrations on the day prior to each exceedance day lower the potential for day-to-day carryover. Conversely, low wind speeds aloft enhance this potential for days within Bin 13 while exceedances on the previous days enhance this potential for Bin 31. Both of these conditions exist for Bin 27.

The potential for contribution from transport ranges from low to moderate (corresponding to low to moderate wind speeds both near the surface and aloft). This potential is lowest for Bins 13 and 27, due primarily to low wind speeds aloft. Conversely, it is highest for Bin 19, due to moderate wind speeds at all levels. Days with a moderate or greater transport potential are identified in Table 3-2 as transport days. All other days are classified as predominantly home grown.

In summary, of the 36 exceedance days considered in this analysis 32 days are predominantly home grown and 4 are characterized by a moderate potential for a contribution from transport. The potential transport days occurred at a average rate of less than one per year during the ten-year period. Considering all possible three-year periods within the analysis period, a maximum of 2 such days occurred within any three-year period.

 

Table 3-5. Summary of classification accuracy for the Charlotte CART Tree.

True Class

 

 

1

2

3

4

C

1

1269

59

1

0

A

2

285

217

25

0

R

3

47

81

91

1

T

4

4

15

10

35

 

 

TABLE 3-6. Summary of classification of exceedance days for Charlotte.

 

 

Bin No.

 

Exceedance Bin

No. of Exceedance Days

Average Maximum Ozone (ppb)

Importance of Recirculation or Day-to-Day Carryover

Potential for Contribution from Transport

Classified as Predominantly "Home Grown"

8

Y

3

138.0

M

LM

Y

13

Y

4

134.0

MH

L

Y

15

N

1

137.0

M

LM

Y

19

Y

4

134.3

LM

M

N

20

Y

2

134.0

LM

LM

Y

27

Y

8

133.0

H

L

Y

29

Y

13

146.1

M

LM

Y

31

Y

1

127.0

H

LM

Y

Key: Y=Yes, N=No, H=High, M=Moderate, L=Low

 

 

Predominantly "Home Grown"

 

Influenced by Transport

No. of Days

32

 

4

Average Maximum Ozone (ppb)

135.2

 

134.3

 

NASHVILLE

The results of the CART analysis for the Nashville area are discussed in this section. This discussion is followed by a detailed analysis of the meteorological and transport characteristics of the days within the exceedance bins

As indicated in the previous section, key classification parameters include wind speeds (near the surface and aloft) and the previous days maximum ozone concentration. In the following summary for the Nashville area, the descriptors low, moderate, and high for surface wind speeds refer to those that are less than 2.5 ms-1, between 2.5 and 4.5 ms-1, and greater than 4.5 ms-1, respectively. For the upper levels, they refer to wind speeds that are less than 4.5 ms-1, between 4.5 and 10.5 ms-1, and greater than 10.5 ms-1, respectively. For ozone, the descriptors refer to concentrations that are less than 85 ppb, between 85 and 105 ppb, and greater than 125 ppb, respectively.

CART Results

The classification tree for Nashville consists of 34 terminal bins (Figure 3-4). Of these, 6 bins are designated exceedance bins. All days on which the maximum ozone concentration at any regulatory site within the Nashville urban area exceeded the current federal ozone standard of 124 ppb were assigned to these bins. The classification accuracy of the CART tree is depicted in Table 3-7. Good classification is achieved within most categories. During the ten-year period included in this analysis, there are 30 exceedance days. All of these are properly classified as having maximum ozone concentrations that are greater than or equal to 125 ppb (Category 4). There are also 36 non-exceedance days misclassified into the exceedance category. Forty-four percent of these have ozone maxima that are between 105 and 125 ppb. Of the remaining misclassified days, more than half are weekend days.

Certain variables were identified by the CART program as being of primary importance in constructing the binary classification tree for the Nashville area. These include yesterdays maximum ozone concentration, average wind speed at the 850 mb level, average surface wind speed between 0700 and 1000 LST, average surface wind speed between 1000 and 1300 LST, average wind speed at the 700 mb level, daily maximum surface temperature, relative humidity at noon, and average height of the 850 mb level.

Review of these parameters and the classification tree (the parameter splits and the average values of the input parameters associated with each terminal bin) indicates that the relationship between the meteorological variables and the ozone concentrations depicted by the CART tree is physically meaningful.

Exceedance Bin Characteristics

Bin 3

Bin 3 contains two exceedance days. It also, however, contains 6 non-exceedance days. According to CART, the days within this bin are characterized by a maximum surface temperature between 27.5 and 32.5 C (81.5 and 90.5 F), low surface wind speeds between 0700 and 1000 LST (less than or equal to 1.57 ms-1), low surface wind speeds between 1000 and 1300 LST (less than or equal to 3.43 ms-1), low relative humidity at noon (less than or equal to 34.5 percent), and an average height of the 850 mb level of greater than 1550 meters. According to CART, this last meteorological feature distinguishes this bin from an adjacent Category 3 bin (Bin 2). The meteorological feature that distinguishes this bin from an adjacent Category 2 bin (Bin 4) is that the relative humidity is lower.

Both the average relative humidity at noon and the average daily maximum surface temperature are lower for this bin than for the other exceedance bins.

Despite the lower temperatures, the data and especially the low surface wind speeds suggest that high ozone concentrations for days within this bin are primarily the result of stagnation conditions near the surface. Low wind speeds aloft and clear skies on the actual exceedance days support the classification of days within this bin as home grown. Recirculation is indicate by the low wind speeds aloft and the relatively high ozone concentrations on the previous day.

Bin 15

Bin 15 contains one exceedance day and no non-exceedance days. According to CART, the day within this bin is distinguished by a maximum surface temperature between 32.5 and 33.6 C (90.5 and 92.5 F), low surface wind speeds between 1000 and 1300 LST (less than or equal to 3.35 ms-1) and between 1300 and 1600 LST (less than or equal to 4.21 ms-1), and low relative humidity at noon (less than or equal to 47.5 percent). Days within this bin are also characterized by average wind speeds at the 850 mb level of greater than 5.43 ms-1. This last feature distinguished this bin from the adjacent Category 2 bin (Bin 14) which requires the wind speed for this level to be less than or equal to 5.43 ms-1. The split on maximum temperature of less than or equal to 33.6 C (92.5 F) distinguishes this bin from the adjacent Category 3 bin (Bin 16).

Compared to other exceedance bins, the average surface wind speed for 0700 to 1700 LST is the lowest as is the ozone concentration on the previous day. Stable conditions are also most pronounced for this bin.

The very low wind speeds suggest that the high ozone concentrations for the day within this bin are primarily the result of stagnation conditions near the surface. Moderately high surface temperatures with clear skies also support this conclusion. However, moderate wind speeds aloft suggest that a contribution from transport is a possibility.

Bin 19

Bin 19 is the most populated of the exceedance bins and contains 17 exceedance days. It also contains 22 non-exceedance days, distributed amongst bins representing the other three ozone classifications. According to CART, the days within this bin are characterized by a maximum surface temperature of greater than 33.6 C (92.5 F), relatively low surface wind speeds between 1000 and 1300 LST (between 0.61 and 3.35 ms-1) and between 1300 and 1600 LST (less than or equal to 4.21 ms-1), surface wind speed between 0700 and 1000 LST of greater than 0.98 ms-1, low relative humidity at noon (less than or equal to 47.5 percent), time of the occurrence of the maximum ozone concentration to be at or before 1600 LST, a lapse rate of greater that -4.45 C between the surface and 850 mb, and average wind speed at 700 mb of less than 4.29 ms-1. This last feature distinguishes this bin from the adjacent Category 2 bin (Bin 20) which requires the wind speed for this level to be greater than 4.29 ms-1. The split on the early morning average wind speed of greater than 0.61 ms-1 distinguishes this bin from the adjacent Category 3 bin (Bin 18).

Compared to other exceedance bins, the average daily maximum temperature is one of the highest as is the average maximum ozone concentration on the previous day.

The data suggest that high ozone concentrations for days within this bin are primarily the result of stagnation conditions near the surface and aloft. Although cloudy conditions complicate the analysis, the high temperatures support the conclusion regarding ozone conducive meteorological conditions. The potential for recirculation is also strongly indicated.

Bin 22

Bin 22 contains 5 exceedance days and 5 non-exceedance days that fall into the Category 3 class. According to CART, the days within this bin are characterized by a maximum surface temperature of greater than 33.6 C (92.5 F), relatively low surface wind speeds for the period 1000 through 1300 LST (between 0.61 and 3.35 ms-1) and for the period 1300 through 1600 LST (less than or equal to 4.21 ms-1), surface wind speeds between 0700 and 1000 LST of greater than 0.98 ms-1, low relative humidity at noon (less than or equal to 47.5 percent), time of the occurrence of the maximum ozone concentration to be at or before 1600 LST, a lapse rate of greater that -4.45 C between the surface and 850 mb, and an average wind speed at 700 mb of greater than 4.99 ms-1. This last feature distinguished this bin from the adjacent Category 3 bin (Bin 21) which requires the wind speed for this level to be between 4.29 and 4.99 ms-1. The requirement that the wind speed for the period 1300 through 1600 LST be less than or equal to 4.21 ms-1 distinguishes this bin from the adjacent Category 2 bin (Bin 23).

Compared to other exceedance bins the average wind speed at 700 mb is greater (although still characterized as moderate).

The data suggest that high ozone days within this bin are primarily homegrown (stagnation) events with high temperatures and low surface wind speeds. The moderate wind speeds aloft suggest some potential for transport.

Bin 25

Bin 25 contains four exceedance days and 3 non-exceedance days distributed evenly amongst the remaining categories. According to CART, the distinguishing characteristics of this bin are average maximum surface temperature of greater than 32.5 C (90.5 F), relatively low wind speeds at the surface between 1000 and 1300 LST (less than or equal to 3.35 ms-1), average wind speed at 850 mb of less than or equal to 1.48 ms-1, and a relative humidity at noon of greater than 52 percent (the average is approximately 55 percent). The low wind speeds at the 850 mb level distinguish the days in Bin 25 from those of the adjacent Category 2 bin (Bin 26).

Compared to the other exceedance bins, the skies tend to be cloudier for the days within this bin, however average maximum temperatures are still relatively high. Average wind speeds at the 850 mb level are the lowest.

The data suggest that high ozone concentrations for days within this bin are primarily the result of stagnation conditions near the surface and aloft and there is no indication of transport but a moderate to high potential for recirculation.

Bin 31

Bin 31 contains only one exceedance day. According to CART, the distinguishing characteristics of this bin are average maximum surface temperature of greater than 32.5 C (90.5 F), a low relative humidity at noon (less than or equal to 30.5 percent), and wind speeds at the surface between 1000 and 1300 LST of greater than 3.35 ms-1 and between 1300 and 1600 LST of greater than 4.86 ms-1. The higher wind speeds for the 1300 LST through 1600 LST period distinguish this bin from the adjacent Category 3 bin (Bin 30). The low relative humidity distinguishes this bin from the other adjacent Class 3 bin (Bin 32).

Compared to the other exceedance bins, wind speeds at the surface and at 850 mb are the highest and (based on temperature lapse rate) the instability is the greatest.

The data suggest that high ozone concentrations for the day within this bin have a moderate probability of being the result of transport. Moderate wind speeds at the surface and aloft support this conclusion.

Summary of Classification of Exceedance Days for Nashville

The results of the CART-based episode classification analysis for the Nashville area are summarized in Table 3-8. For each bin, the importance of day-to-day carryover or recirculation of ozone was estimated primarily based on the previous days ozone concentration and wind speed aloft.7 However, the other characteristics of days within the bin, as described in the preceding text, were also considered. The potential for contribution from transport was estimated primarily based on surface and upper-air wind speeds.8 Again, other factors as described above were used to confirm the characterization.

For the exceedance days considered in this analysis, day-to-day carryover or recirculation ranges from moderate to high importance. The criteria for recirculation correspond to high ozone concentrations on the previous day (greater than or equal to 125 ppb) and low wind speeds aloft. For Bin 15, moderate wind speeds aloft coupled with low ozone concentrations on the previous day lower the potential for recirculation. Due to lower wind speeds aloft and low ozone on the previous day, the potential for recirculation for the days in Bin 3 is moderate. Moderate ozone concentrations on the previous day coupled with moderate wind speeds aloft increase the potential for recirculation for Bins 22 and 31 and, therefore, these bins were also classified as having a moderate potential for recirculation. For Bins 19 and 25, ozone concentration on the previous day is moderate and wind speeds aloft low, therefore, days within these bins were classified as having a moderate to high potential for recirculation.

Due to low wind speeds at the surface and aloft, the potential for contribution from transport is low for most of the exceedance days. Twenty three of the 30 exceedance days had these meteorological characteristics. Two bins, Bins 15 and 22 have a low to moderate potential for transport and are characterized by low surface wind speeds and moderate wind speeds aloft. One bin, Bin 31, which contained only one exceedance day, had moderate wind speeds at the surface and aloft, indicating potential transport. Consequently, only one of the 30 exceedance days studied in this analysis can be characterized as having a moderate potential for contribution from transport.

All exceedance days that occurred during the 1995 Nashville Southern Oxidant Study (SOS) were classified as home grown in this analysis. This is consistent with the SOS data results.

 

Table 3-7. Summary of classification accuracy for the Nashville CART Tree.

True Class

 

 

1

2

3

4

C

1

1229

27

0

0

A

2

401

182

24

0

R

3

77

53

81

0

T

4

14

6

16

30

 

 

TABLE 3-8. Summary of classification of exceedance days for Nashville.

 

 

Bin No.

 

Exceedance Bin

No. of Exceedance Days

Average Maximum Ozone (ppb)

Importance of Recirculation or Day-to-Day Carryover

Potential for Contribution from Transport

Classified as Predominantly "Home Grown"

3

Y

2

136.0

MH

L

Y

15

Y

1

142.0

LM

LM

Y

19

Y

17

135.5

H

L

Y

22

Y

5

134.0

M

LM

Y

25

Y

4

133.0

MH

L

Y

31

Y

1

125.0

M

M

N

Key: Y=Yes, N=No, H=High, M=Moderate, L=Low

 

 

Predominantly "Home Grown"

 

Influenced by Transport

No. of Days

29

 

1

Average Maximum Ozone (ppb)

136.1

 

125.0

 

4 SUMMARY OF FINDINGS

In summary, the results of the ozone episode classification analysis for the cities of Atlanta, Birmingham, Charlotte, and Nashville using the Classification and Regression Tree (CART) analysis technique as the primary classification tool indicate that:

The combined results for all four areas suggest that regional-scale transport of ozone does not play a predominant role in the occurrence of ozone exceedance days within these urban areas. Given the variable number of exceedance days for each area and the low incidence of transport, these results also suggest that ozone transport among these areas does not frequently influence exceedances of the ozone standard. A moderate potential for contribution from transport is indicated for certain exceedance days within the Atlanta, Charlotte, and Nashville areas (these days comprise 6, 11, and 3 percent, respectively, of all days included in this analysis). In all cases, the potential transport days occurred at a average rate of less than one per year during the ten-year period. Considering all possible three-year periods within the analysis period, the number of days that occurred within any three-year period did not exceed three. Thus, transport days alone (as identified by this analysis) would not have caused any of the areas to be in violation of the current federal ozone standard.

 

 

 


Submit your comments, feedback, questions, and ideas pertaining this page. Your input will be automatically added to the existing annotations. In order to add a new comment, you must be registered with the OTAG/AQA Peoples Page.