Ozone as a Function of Local Wind Direction and Wind Speed: Evidence of Local and Regional Transport


Rudolf B. Husar and Wandrille P. Renard

Center for Air Pollution Impact and Trend Analysis (CAPITA)

Washington University

St. Louis, MO 63130-4899


May 7, 1997




The main factors that influence the variability of ozone are the wind direction and wind speed, assuming that the emissions and the chemistry are unchanged. A simple way to examine the role of transport (wind direction and wind speed) is to classify the existing ozone concentration data into wind direction and wind speed ranges. Given long enough sampling record, say, ten years, the dependence of ozone on transport can be extracted from the climatological record for ozone and winds.

A full quantification of ozone transport in an unambiguous and robust way has so far eluded the analysts in the Air Quality Analysis Workgroup. However, at this time there is a multiplicity of evidence that ozone has both local impact and it is also transported regionally and causes impact 500-1000 km from the source.

The current analysis can be viewed as a complement to backtrajectory analysis (Poirot and Wishinsky, 1996), forward trajectory and regional impact analysis (Schichtel and Husar, 1996) and analysis of aircraft and surface observations in the Northeast (Blumenthal et al., 1997).

This report was prepared to support the deliberations of the OTAG Air Quality Analysis workgroup. The specific purpose of this report is to examine the dependence of ozone concentration on local and regional ozone transport.

OTAG Mission and Goals

The mission of OTAG is to identify control strategies and implementation options for the reduction of regional ozone over the eastern US The operational goals of OTAG are stated as (1) A general reduction in ozone and ozone precursors aloft throughout the OTAG region and (2) a reduction of ozone and ozone precursors at the boundaries of nonattainment areas.

Policy-Relevant and Scientific Results

It is suggested that the directionally and wind speed sorted ozone data can be used for several purposes that are of interest to OTAG.

  1. Source identification. The location of an ozone source can be "triangulated" using directional sorting of ozone data. In case of a source, such as St. Louis, the elevated ozone concentration "downwind". If the concentrations at monitoring sites surrounding a source "point" to a common location as a source of elevated ozone, then it can be stated that the particular source "causes" the elevated concentrations in its vicinity.
  2. Imported vs. "homegrown" ozone apportionment. The directionally sorted ozone concentrations can be utilized to determine the magnitude of regional ozone at a given site by eliminating those measured concentrations that are "downwind" from the local source.
  3. Transported ozone estimation. During high wind speeds elevated ozone concentrations can only be attributed to transported ozone. Hence, the technique can provide an estimate of transported ozone.

Data Sources and Processing

Data Sources and Quality Control

The ozone data used in this report were collected from multiple sources:


Data Set

Supplying Organization









Eulerian Model Evaluation and Field Study



Southern Oxidant Study

1993, 1995


Lake Michigan Air Directors Consortium

1991 (88, 93, 95)


State of Georgia

1988, 91, 93, 95


State of North Carolina



Data from each network were extracted and combined into a single integrated data set. The details of the data sources and quality control procedures are discussed in the report "Preparation of Ozone Files for Data Analysis".

The first examination of average daily maximum ozone maps has revealed anomalous ozone "holes" and peaks at unexpected locations. For those sites the hourly and daily maximum ozone values were re-examined for possible inconsistencies. Sudden systematic changes in the ozone concentrations, as well as major deviation from neighboring sites were the main clues for anomalous behavior. As a result of this quality control process, 6 out of 709 monitoring sites were discarded. The remaining data were used in all the subsequent computations exactly as submitted by the networks.

Data Processing Procedures

The data processing was conducted in the following major steps below:

  1. Data from individual networks were quality controlled and formatted uniformly.
  2. The hourly ozone data from all the networks were combined into a single database.
  3. The daily maximum (1-hour average) ozone was extracted from the hourly data.
  4. For each monitoring station the average, percentiles and exceedances of daily maximum ozone was computed.
  5. The results for all stations were contoured and plotted on maps and for easy presentation.

In this analysis the ozone data for the 1986-1995 were merged with wind direction and wind speed data from meteorological monitoring sites. For every ozone monitoring site the nearest meteorological monitoring site was identified and assigned to the ozone site. If the closest meteorological site did not have direction and wind speed, then the next closest meteorological site was selected.

The wind direction and wind speed was obtained from the National Weather Service synoptic monitoring network, consisting of about 300 sites in the conterminous US. The wind direction and wind speed represent the surface observations at local noon.

The ozone concentrations have been sorted and averaged for specific wind direction and speed ranges for every monitoring site. The average ozone concentration was computed for each wind direction range in 45 increments. The first directional wind was between 0-45, i.e. when the wind blew from north or northeast. This resulted in 8 wind directional concentration bins. The average concentrations for each directional bin was further classified by wind speed, ranging between 0-2, 2-4, 4-6, 6-8 m/sec increments. Thus, there were eight directional and four wind speed bins, yielding a total of 32 concentration bins. For presentation of the results, some of these concentration bins were lumped in order to increase the statistical robustness.

Ozone as a Function of Local Surface Wind Direction and Wind Speed

The results of the analysis are presented in three different forms:

  1. Maps of average ozone concentrations for specific wind direction and wind speed. These yield a spatial pattern of ozone in the OTAG domain for different wind conditions.
  2. Ozone roses, i.e. average ozone concentration as a function of wind direction for specific subregions. These are useful "pointers" toward the ozone source areas.
  3. Charts of average concentrations as a function of wind speed and wind direction at specific locations. These charts are helpful in illustrating the role of transport.

The maps of average ozone concentration for four wind directional sectors is shown Figure 3 a,b,c,d,e. The first four show the concentration for wind directional quadrants. The last Figure 3e depicts the average concentration for all wind directions.

The directionally classified ozone concentrations indicate that at low wind speeds (<3 m/sec) (Figure 3 a-e) ozone levels do not vary substantially with wind direction. The highest concentrations are found in the northeastern urban corridor, as well as in Atlanta, Dallas, Houston, St. Louis, and over the Ohio River Valley. The directionality of ozone concentrations in the vicinity of urban areas is noticeable but it is not significant. At the wind speed of 2 m/sec, the transport distance of an air mass is about 200 km per day, as indicated by the length of the arrows, placed over several OTAG locations. The time period of one day was chosen since it represents the approximate time between the precursor emissions and the ozone removal. The length of the arrows can be used as a guide to the eye. It is evident, that at low and meandering wind speeds, local emissions are the main causal factors of the observed ozone concentrations and that the wind direction is virtually irrelevant.

The directionally classified concentrations at intermediate wind speeds (3-6 m/sec) indicate substantial differences between the ozone maps for the four directional quadrants (Figure 4 a,b,c,d,e). For the wind directions during northerly quadrants (0-90 and 270-360 ) (Figs 4a and c), the concentrations are low throughout the northern belt of OTAG. Evidently, northerly winds bring low ozone concentrations to Minnesota, Wisconsin, Michigan, and New England. Northerly winds are also associated with higher ozone concentrations in the southern belt of OTAG from Texas through Georgia.

When the winds are generally from the south (90-270), the northern belt of OTAG, Minnesota to Maine is experiencing higher ozone concentrations, particularly in the mid-section from Michigan through New York. At the same time, during the southerly winds the ozone concentration in the southern belt from Texas to Georgia is low. Clearly, southerly winds bring low ozone to the Gulf states.

Within the states adjacent to the Ohio River Valley, northerly winds (0-90 and 270-360) are associated with higher ozone levels in Kentucky, Tennessee and West Virginia, i.e. just south of the Ohio River Valley. On the other hand, southerly winds (90-270) are associated with higher ozone concentrations in the belt stretching from Illinois through Pennsylvania, just north of the Ohio Valley.

At moderate wind speeds, the concentration pattern over OTAG indicates features of a tidal wave of ozone that appear to drift in the direction of the winds. The directionality of ozone concentrations in the vicinity of urban areas is more noticeable than at low wind speeds. Most major urban areas appear to have higher concentrations downwind, regardless of which direction the wind is blowing. At the wind speed of 5 m/sec, the transport distance of an air mass is about 500 km per day, as indicated by the length of the arrows in Figures 4 a-d. Based on the directional alignment of ozone, it is suggested that at moderate wind speeds, local emissions as well as regional transport are contributors to elevated ozone concentrations.

At high wind speeds (>6 m/sec), the directionally classified concentrations indicate major regional concentration differences depending on wind direction (Figure 5 a,b,c,d,e). Qualitatively, the pattern is similar to the moderate wind speeds: during northerly winds the northern belt of OTAG has low ozone, while the southern belt Texas, Georgia has higher values. However, the directional displacement of ozone is on the order of 1000 km during high wind speeds. It appears that at >6 m/sec most of the ozone is blown into the "downwind" quadrants of OTAG while the other three quadrants remain at low levels.

On the other hand, during strong southerly winds (90-270), the entire southern belt is virtually ozone free, but the northerly states from Illinois through New England experience elevated concentrations. The concentrations of ozone in northeastern quadrant of OTAG are particularly high during strong southwesterly winds (180-270).

During high wind speeds the concentrations in urban areas do not differ significantly from the regional background. Notable exceptions are the northeastern corridor, where the urban concentrations are elevated. At the wind speed of 8 m/sec, the transport distance of an air mass is about 1000 km per day, which is comparable to the size of the OTAG domain.

The examination of ozone concentrations at wind speeds 6 m/sec or above, appears to be of particular interest to OTAG. At such wind speeds (50 km/hour) all the measured ozone concentrations could be called regional by the following reasoning. Even during the photochemically active midday hours, it takes 1-2 hours of irradiation in order to create the bulk of ozone from local precursor emissions. During the ozone formation time, the atmosphere moves the emissions 50-100 km from their source. Conversely, any ozone that is measured at a location at high speeds must have been emitted at least 50-100 km upwind. One could say, therefore, that "homegrown" ozone is non existent at high speeds; virtually the entire ozone mass must be regionally transported ozone. In this line of reasoning, the distinction between "homegrown" and regional was made based on the radius of about 50-100 km from the source.

Examination of the directionally sorted ozone concentrations is particularly instructive when viewed through animations. The four following animated maps represent the wind directionally sorted average concentrations at low (<3 m/sec), moderate (3-6 m/sec), high (>6 m/sec), and all wind speeds. Additional animations illustrate the ozone pattern when the wind is blowing in the same direction two days in a row and when only considering wind speed (independent of wind direction). (download the movies).

The obvious limitations of this analysis includes the use of surface winds, opposed to mean transport winds, and location differences between meteorological and ozone sites.

Directional Ozone Roses

One of the ways to examine the directionality of ozone transport is through ozone pollution roses. In the current usage an ozone pollution rose is defined as an average ozone concentration when the wind blows from a given direction. It does not consider the frequency of occurrence of different wind directions. The purpose of such a analysis is to evaluate the magnitude of the ozone concentrations that arrive at receptor site from different directions. The wind directional sectors that show high values of the ozone rose provide "fingers" that point toward the source areas of high ozone concentrations.

Figure 7 shows ozone concentration roses at nine selected subregions of OTAG. The ozone roses display the concentrations that are above 50 ppb. The ozone rose for each subregion was averaged using monitoring sites in a 100 km rectangular box.

At Philadelphia, when the wind blows from the north the average concentration is about 60 ppb. Winds from the southwest are associated with over 80 ppb average ozone concentration. This points to the southwest as the source of high ozone in Philadelphia.

In Atlanta, southerly winds are associated with average ozone of 60 ppb. On the other hand, when the wind blows from the north the average concentration is over 80 ppb. The apparent "source direction" of high ozone in Atlanta is to the north.

In St. Louis, the ozone concentrations are about 55 ppb during winds from the west and northwest, and increase to about 70 ppb when the winds are from the east, south, and southwest. Evidently, St. Louis receives elevated ozone from the east and south.

The directionality of concentrations at western Michigan and Detroit is such that concentrations over 50 ppb occur during southerly winds. On the other hand, in Houston, the average ozone concentration is over 70 ppb when the winds are northerly, and less than 40 ppb during southerly winds.



Figure 3 a.


Figure 3b


Figure 3c


Figure 3d


Figure 3e


Figure 4a


Figure 4b


Figure 4c


Figure 4d


Figure 4e


Figure 5a


Figure 5b


Figure 5c


Figure 5d


Figure 5e


Figure 6a


Figure 6b


Figure 6c


Figure 6d


Figure 6e



Figure 7


Distinction between Local and Regional Ozone Using Wind Speed and Direction

The ozone concentration at a given location and time is contributed by biogenic sources (tropospheric background), anthropogenic sources that are more than 100-200 km from the receptor (regional ozone), and "homegrown" ozone from sources that are <100-200 km from the receptor (local ozone).

The magnitude of the tropospheric background ozone (30-40 ppb) has been established through monitoring data at locations removed from anthropogenic sources. The other known quantity is the total ozone concentration which is the sum of biogenic, regional, and local contributions. A reliable apportionment of the total measured ozone among the three components has eluded analysts over the past decades. The analysis below is a crude attempt to estimate the role of regional and local ozone using the ozone concentration dependency on wind speed and direction. The analysis is based on a simple premise that if the measured ozone concentration declines steadily with increasing wind speed, i.e. ventilation, then the ozone is largely "homegrown" contributed by local sources. On the other hand, if the ozone concentration does not decline with wind speed, then the ozone is attributable to distant sources, i.e. regional sources.

It is instructive to examine the wind speed dependence of ozone using a simple one dimensional transport model (Figure 7a {not done leave placeholder}). The pollutant emissions are confined to a mixing layer of H[m]. Within the mixing layer the unidirectional wind speed is U[m/s] and carries a background concentration C0[g/m3] into a source area. The source area itself has an emission density of Q[g/m2,s] as well as a length of L[m]. Assuming that the local emissions are mixed instantaneously, the concentration, C[g/m3], averaged over the source region can be estimated by the expression:

C = C0 + QL/UH

The second term on the right side represents the local contribution and it is proportional to the source characteristics (QL) and it is inversely proportional to the ventilation coefficient (UH). The dependence of the local contribution on wind speed is inverse, as illustrated in Figure 7b. The concentration, C, is highest at low wind speeds because the pollutants accumulate due to poor ventilation. With increasing wind speed, concentrations asymptotically approach the regional background concentrations due to the rapid dilution of the local contributions.

In the analysis below, the above simple model is used to interpret the measured ozone concentrations as being of local or regional in origin. Strongly declining concentrations with wind speed will be interpreted as evidence of local source contributions, since higher wind speeds cause increasing dilution of local contributions. If the concentration is found to be constant with wind speed, then it taken as evidence that the local contribution is not significant, hence regional transport dominates.

Ozone Concentration as a Function of Wind Speed

The results of the wind speed segregated ozone concentrations are shown in Figure 8 a-m. Each figure shows the average ozone concentration at 1 (0-2), 3 (2-4), 5 (4-6), and 7 (6-8) m/s. The data are further stratified by 4 wind directional quadrants from 0-90 , through 270-360 . A fifth line in the chart represents the wind speed dependence of ozone, regardless of the wind direction, incorporating the directional frequency of ozone occurrences. It needs to be noted, that ten years, 1986-1995, of ozone and meteorological data were used in the statistics. Nevertheless, for some classification bins, particularly at high wind speeds was limited. In order for a station to qualify, at least ten days of data was required for a wind direction/wind speed bin.

Southwestern metropolitan areas. The wind speed charts for the southwestern metropolitan areas show strongly declining concentrations with increasing wind speeds. This is indicative that local sources are the dominant contributors to the average ambient ozone concentrations . For Dallas-Ft. Worth (Figure 8a) the subregional average ozone concentrations are about 80 ppb at 1 m/s, declining to about 55 ppb at 7m/s wind speed. Assuming that the ozone at the high wind speed is all regional, Dallas metropolitan area contributes about 25 ppb (80-55) to the regional background of 55 ppb. The wind speed dependence of ozone at Dallas-Ft. Worth does not depend on wind direction since; the lines for all four wind directional sectors overlap. In this regard, Dallas-Ft. Worth is unique among the metropolitan areas.

Data for Houston, TX show a remarkably strong decline of ozone concentration with wind speed. During stagnating conditions (0-2 m/s) the average concentration at the subregional monitoring sites is 85 ppb, and there is a near linear decline to about 35 ppb at high, 7 m/s wind speeds. Taking the regional ozone entering the city as 35 ppb, the Houston metropolitan area contributes 50 ppb (85-35) to the regional background. The wind directional dependence of ozone clearly shows, that during northerly winds (270-360 , and 0-90 ) the regional background ozone levels are about 20 ppb higher than during the southerly winds from the Gulf of Mexico. In fact, during strong northerly winds, the regional ozone (about 50 ppb) is comparable in magnitude to regional ozone estimate at Dallas-Ft. Worth.

Southeastern metropolitan areas. Southeastern metropolitan areas show both local and regional contributions. At Birmingham, AL, low wind speeds are associated with moderate ozone concentration of about 75 ppb, which decline to 50 ppb at 7 m/s. At low wind speeds the concentrations are almost independent on wind direction. However, at high wind speed, southerly winds are associated with 40 ppb, while northerly winds carry 60 ppb of ozone. This pattern is similar to Houston and indicates that northerly winds carry regional ozone that is about 20 ppb higher than the tropospheric background of 30-40 ppb.

Atlanta, GA also shows unique wind speed and direction dependence of ozone. The average concentration at low wind speeds is about 80 ppb and declines to about 55 ppb at 7 m/s. The unique feature for Atlanta is that during southerly winds the concentration decline is roughly linear with wind speed. On the other hand, winds from the northern sectors show constant ozone concentration up to 4 m/s, followed by decline at higher speeds. As a result, at high wind speeds winds from the northwest (270-360 ) bring ozone levels of 65 ppb, while southeasterly winds form the Atlantic bring <50 ppb ozone. It is evident, that in Atlanta, both the ozone concentration at high wind speeds, as well as the shape of wind speed dependence differs for northerly and southerly winds.

The ozone level at Charlotte, NC is about 75 ppb at low wind speeds, and about 55 ppb at high wind speeds. Overall, the concentration decline with wind speed is moderate, except for strong southeasterly winds from the Atlantic when the average ozone concentration drops to 35 ppb. At Charlotte, NC the ozone levels are highest when the winds blow either from the northwest or from the southeast.

The average ozone concentrations at Nashville, TN is about 65 ppb at low wind speeds and declines to about 50 ppb at about 7 m/s, yielding a high-low speed difference of only 15 ppb. Evidently, the Nashville metropolitan area contributes less local ozone than the other major southern metropolitan areas. At high wind speeds the ozone concentrations of northerly winds are about 10 ppb higher than southerly winds.

West-central metropolitan areas. The west-central metropolitan areas under consideration include St. Louis, MO, Chicago, IL, and Detroit, MI. In St. Louis, MO, the ozone level at low wind speed is 80 ppb, at high speed is 50 ppb, with an excess of 30 ppb (80-50) at low wind speeds. The unique feature of St. Louis is that at low wind speeds there is a remarkable directionality of ozone concentrations. In general, southerly winds are associated with higher ozone levels than northerly directions. Based on strong wind speed dependence of average ozone, the St. Louis metropolitan area has major local impact on itself.

The Chicago, IL subregion has remarkably low average ozone concentration of 62 ppb at < 2 m/s, and declines to about 50 ppb at high speeds. The unique feature of Chicago is that winds from the south (90-270 ) are associated with about 60 ppb of ozone, regardless of the wind speed, while winds from the northerly sectors show substantial decrease of ozone with wind speed. In this case, the evidence suggests that during southerly winds the regional ozone concentration of 60 ppb dominates the Chicago subregion, while during northerly winds, local emissions contribute 20-30 ppb to the background ozone.

Detroit, MI exhibits an interesting pattern in that the average ozone concentration changes only 7 ppb from low wind speeds (62 ppb) to high wind speed (55 ppb). This weak dependence on the wind speed arises from the compensating effects of different wind directions. During southerly winds the ozone concentration increases slightly from 60 ppb at low wind speeds to 65 ppb at high wind speeds. On the other hand, at northerly winds, the concentrations decline from over 60 to 45 ppb. It is evident therefore, that the average ozone concentration at Detroit is dominated by the regional background concentration of 60 ppb during southerly winds, while during northerly winds local emissions appear to contribute about 15 ppb on top of a regional background of about 45 ppb.

Northeastern corridor. The populated northeastern corridor is known for high density of ozone exceedances, high emission density of ozone precursors, as well as evidence that regional scale ozone transport into the region is significant.

Washington, DC shows moderate decline of 15 ppb (80-65) ozone concentration with increasing wind speed. Southerly winds show somewhat higher concentration then northerly directions. The lowest ozone levels are observed when the winds blow from the northwesterly quadrant. It is evident, that regional ozone (65 ppb) is a large fraction of the total average ozone (80 ppb) in the Washington metropolitan area.

Philadelphia, PA and New York City metropolitan areas show qualitatively similar dependence on wind speed and wind direction. The average concentration at high and low wind speeds differs only by about 5 ppb (70-65). At low wind speeds the concentrations are roughly independent on wind speeds. However, at high wind speeds southwesterly winds carry about 80 ppb of ozone, while swift northeasterly winds have ozone concentrations near the tropospheric background. It is beyond doubt, that southeasterly winds in Philadelphia-New York City corridor transport ozone into these metropolitan areas with an average concentration of 80 ppb. The wind speed dependence of ozone during northeasterly events from the Atlantic show a strong decline with wind speed, which is characteristic for local contributions. Hence, during northeastern winds in the summer, local contributions from the New York and Philadelphia metropolitan areas dominate the ozone concentrations. On the other hand, during strong southwesterly winds, virtually all the ozone concentrations are contributed by regionally transported ozone at 80 ppb concentration.

The wind speed and directional dependence of ozone in the Philadelphia-New York corridor deserves further consideration. During southwesterly winds, the ozone concentration actually increases with increasing wind speed up to 5 m/s (500 km per day). A possible explanation of ozone increase with wind speed is that ozone from the west-southwest is transport from major distant sources where ozone removal between the source and receptor is significant. At higher wind speeds the transport time is reduced which can result in less removal and increased concentration. The other issue with regard to the transport in the Northeast is that the wind direction derived from the surface winds may not be representative of the wind direction for the bulk ozone transport within the entire boundary layer. In fact, Blumenthal et al. 1997, suggested that while surface winds tend to be southwesterly direction, the main transport winds in the 200-800 m elevation tend to be from the west.

The Boston, MA metropolitan area shows virtually no dependence of ozone concentration on wind speed, except during northeasterly winds. The lack of wind speed dependence clearly indicates that the average concentration in Boston is dominated by transport and that the local contributions to the average are virtually undetectable. Directionally, southwesterly winds are the highest at 70 ppb, and northeasterly transport brings lowest ozone concentrations at about 45 ppb.











Figure 8 a - m


Figure 9 shows the relative change of ozone concentration among different cities. All concentrations are normalized to the values at 1 m/s. For southern urban areas, where episodes are caused by local stagnation, ozone levels decline rapidly with increasing wind speed. In northern cities, more heavily influenced by transport, ozone levels decrease much less rapidly, with increasing wind speeds.


Figure 9



This analysis is too tentative to warrant conclusions at this time. Rather, it demands an extensive disclaimer.

The above source identification and source apportionment techniques have not been fully evaluated in this report. It is clear, however, that the wind direction and wind speed sorting has a potential to yield observation-based evidence regarding ozone transport. Substantial future efforts need to be invested in evaluating the consistency of this method with the other approaches that attempt to illuminate and quantify ozone transport.