AIR QUALITY TRENDS IN ILLINOIS


From The Changing Illinois Environment: Critical Trends,Volume 2: Air Resources, Technical Report of the Critical Trends Assessment Project

INTRODUCTION

Purpose

There are three main purposes for including a component on air quality in a comprehensive assessment of the Illinois environment. The first purpose is simply to document the current status of Illinois air quality and any recent trends. Specifically, we will look for changes (trends) in pollutant concentrations over time within specific geographic regions, as well as statewide, and also show spatial variations of pollutant concentrations in the Chicago area, the only area of the state where there are enough sampling sites to permit plots of spatial variability.

These analyses will reveal whether concentrations of specific pollutants in Illinois air are getting better, or getting worse, or staying the same. They may also indicate which areas of the state experience the highest and lowest pollutant concentrations, and whether such local or regional conditions are improving or worsening.

The second purpose is to provide the information necessary for assessments of human and ecosystem exposure to airborne pollutants. These assessments will appear in other reports in this series.

The final purpose is to identify gaps and needs in Illinois air quality monitoring and research.

Scope

This analysis, for the most part, relies on data generated by routine air quality measurements carried out by the Illinois Environmental Protection Agency (IEPA), the state agency charged with monitoring compliance with state and national air quality standards. It should be noted that IEPA's main purpose in making these measurements is to monitor compliance, not to document the state's air quality. Thus, to some extent, this examination of IEPA data for descriptions of past and current air quality goes beyond the purpose for which the data were originally collected. Consequently, the goal of characterizing air quality over the whole state may not be completely attainable.

Pollutants Analyzed

The air pollutants to be examined for temporal and spatial trends include, first of all, the seven criteria pollutants, i.e., those for which national or state air quality standards have been set. These include four gaseous pollutants: ozone (O3), sulfur dioxide (SO2), nitrogen dioxide (NO2), and carbon monoxide (CO). There are also standards for three pollutants that occur as particles: These are lead (Pb); total suspended particulate matter (TSP), for which there is only a state standard; and particulate matter with aerodynamic diameters of 10 micrometers (mm) or less (PM10), for which there is only a national standard.

In addition, we will examine measured concentrations of several additional pollutants for which standards have not been set. This group includes sulfate and nitrate ions and the metals arsenic (As), cadmium (Cd), chromium (Cr), iron (Fe), manganese (Mn), and nickel (Ni). These pollutants all occur as particles in the atmosphere, and they are measured by chemically analyzing the same filters used to collect the TSP samples. Several additional metals, such as beryllium, copper, selenium, and vanadium, are currently being measured or have been measured in the past by IEPA. These additional metals are not included in this assessment because their concentrations are mostly smaller than the detection limit.

Finally, we will summarize occasional (non-routine) measurements of organic compounds. These measurements were made by IEPA or published in the scientific literature.

Period of Record

The Illinois data to be analyzed represent, at the most, the years between 1978 and 1990. Some pollutants were measured only during the later years of this time period.

Data Quality

Methods of air pollutant sampling and analysis used by IEPA conform to U.S. EPA standards. However, it should be noted that artifacts can occur on certain types of filters possibly used by IEPA for high volume sampling during the data period. Positive sulfate artifacts (formation of sulfate on the filter from gaseous pre-cursors) have been reported (Appel et al., 1984). Both positive and negative (evaporative loss) artifacts can occur for nitrate (ibid.).

METHODS

Data Source

The data used in this report were taken from annual summary reports published by IEPA for the years 1978-1990 (IEPA, 1979-1991). Questionable values were checked with IEPA personnel (Swinford, personal communications, 1992 and 1993) and corrected if necessary.

Creation of Computer Files

Data were entered by hand into computer spreadsheet and database files from IEPA reports. Sampling site locations were specified by latitude and longitude. Spreadsheet data files were also converted to Geographic Information System (GIS) files using existing data conversion software. Site locations in the Chicago area were verified by comparison of locations plotted on GIS maps against road atlas maps, and corrected when necessary.

Statistical Methods

Box Plots. Box plots were used to convey information about the distribution of a particular pollutant and its time-averaged concentrations (e.g., annual mean, highest 24-hr average, or highest 1-hr average) over sampling sites in a specific geographical region for a specific year. For a given year, a box plot shows the 10th, 25th, 50th, 75th, and 90th percentiles, respectively, as the low "whisker", the bottom of the "box", the line across the box, the top of the box, and the upper whisker. Individual values outside the 10th and 90th percentiles were also plotted as individual points. Thus, a plot of a series of boxes representing a series of years visually shows how various percentiles of the distribution of pollutant concentration change over time. In addition to showing plots to give a visual impression of time trends, we have assessed the statistical significance, or lack thereof, of changes in pollutant concentrations over time.

Trend Analysis. Time sequences of pollutant concentrations were tested for statistical significance of time trends over the entire period of record using the nonparametric Spearman Rank Correlation Coefficient as described by Snedecor and Cochran (1980). This is the same trend test used by IEPA to detect the statistical significance of time trends in pollutant concentrations. Please note that the critical values of the rank correlation coefficient for significance at various confidence levels given by IEPA (e.g., IEPA, 1991) differ somewhat from critical values given by tables A11(i) and A11(ii) in the appendices of Snedecor and Cochran (1980), which were used in this study. The test was applied to the series of medians of the annual distributions over sites within each geographic area.

Where significant time trends were found, average annual percent changes were calculated by dividing the slope by the concentration computed for 1984 (the middle year of the 1978- 1990 series). The 1984 concentration was computed from the linear regression line fitted to the time series of median regional concentrations.

Regional Differences. One-way analysis of variance was used to test for differences in mean pollutant concentrations between independent regions; i.e., the Chicago area, the Metro East area, and the remainder of the state. Any data sets that failed the Bartlett's test for homogeneity of variance were also analyzed by the distribution-free Kruskal-Wallace one- way analysis of variance. Differences between individual regions were evaluated from pairwise comparison probabilities computed from the Tukey HSD multiple comparisons procedure. These tests were all carried out using SYSTAT (Wilkinson, 1990).

TIME TRENDS IN POLLUTANT CONCENTRATIONS

Trends in pollutant concentrations over time are the major emphasis of this part of the Air Resources volume. Table 1 presents computed median pollutant concentrations from the distribution of site values for each geographical region and year. Table 2 presents results of statistical tests for significance of time trend for all pollutants examined. Time trends for each pollutant are graphed and the text discusses the results of the related statistical tests in table 2. The graphs show concentrations at several percentile points on the distribution of particular pollutant concentration statistics (i.e., annual means or 24-hr maxima) over all sampling sites within a given geographic area for each of a series of years, and how these percentiles of concentration change over the period 1978-1990. The box plot is a convenient graphical device to use for this purpose, because it shows several concentration percentiles simultaneously. The concentrations plotted in a single graph are averages over a particular time period, such as one hour, 24 hours, or an entire year. For the criteria pollutants, these averaging times correspond to those for which the standards are written for each particular pollutant, so they vary from one pollutant to the next, and a given pollutant may have two or more graphs, corresponding to two or more averaging times.

It will become apparent later that the number of sampling sites for a given pollutant in any geographical area changed from year to year as new sites were installed or previously used sites were taken out of service. In recent years the number of sampling sites has dropped markedly for most pollutants. This suggests an alternative approach to the analysis of time trends: analysis of time trends only at sites with long-term records. This approach was rejected in favor of the analysis of time trends in regional medians because very few sites had long-term records, and such sites could not be inferred to represent broad geographic areas.

Data used in the following statistical tests are shown in table 1.

Results of the statistical tests for the significance of trends are shown in table 2, and the time sequences of box plots in figures 1-30. Each figure shows results for up to four geographical areas, depending on whether the number of sampling sites in the area was sufficient to justify a separate plot. Each figure shows results for the entire state and the Chicago area, which is the portion of the U.S. EPA's Air Quality Control Region (AQCR) 67 in Illinois. Many figures also show results for the Metro East area on the Illinois side of the Mississippi River across from St. Louis. This is the Illinois portion of AQCR 70. The fourth geographic region shown is the remainder of the state, excluding the Chicago area and the Metro East area (if shown separately). If the Metro East area is not shown separately, then the region labeled as "remainder of state" also includes any data from the Metro East area.

Criteria Pollutants

Carbon Monoxide (CO). Results for CO are shown in table 2 and figures 1 and 2. The table shows trends toward decreasing concentrations significant at the 5 percent level, or better, for both 1-hr (figure 1) and 8-hr (figure 2) maxima during the 1979-1990 period, in all three geographic areas with sufficient data to test. The overall linear trends range from -2.7 percent to -5.3 percent. Figure 1 suggests generally higher and slightly increasing 1-hr maximum concentrations at all percentile levels represented by the box plots during the 1978-1984 period, followed by lower and decreasing concentrations from 1985 onward. This pattern appears in all three geographic areas shown in the figure. The 1-hr CO standard, 35 parts per million (ppm), appears not to have been exceeded anywhere in the state during the 13 years of record.

On the other hand, the 8-hr maximum concentrations of CO (figure 2) exceeded the standard, 9 ppm, repeatedly until very late in the 1979-1990 period. Figure 2 shows a pattern of higher concentrations at all percentiles during the 1978- 1984 period, followed by lower and decreasing concentrations from 1985-1990, in all geographic areas.

The statistical comparison of CO concentrations between geographical regions found a significant difference (5 percent) between the Chicago area and the remainder of the state for the 1-hr (figure 1), but not for the 8-hr (figure 2) averaging time. For both averaging times, however, it appears that early in the 1979-1990 period the Chicago area contributed the highest individual concentrations, whereas the highest values during the later years of the period came from elsewhere in the state.

For both CO averaging times, the number of sampling sites (N) remained relatively constant, with about 15 sites in the entire state, 6 to 11 sites in the Chicago area, and 6 or 7 sites elsewhere.

Lead (Pb). Airborne Pb concentrations decreased dramatically (figure 3) during the 1979-1990 period, due in part to the phase-out of leaded gasoline, beginning in about 1975 (IEPA, 1991). Table 2 shows decreasing linear trends between -12.6 percent and - 24.8 percent per year in the four geographical regions. While the trends are significant at the 1 percent level in all four regions, the rate of decrease in the Metro East area is considerably smaller than the others. Figure 3 graphically shows the marked drops in all the concentrations corresponding to the percentiles depicted by the box plots.

There has clearly also been a drop in the highest annual Pb concentration in each region, as depicted by the uppermost open circle. This drop-off was quite dramatic in the entire state, where numerous concentrations exceeded 1 microgram per cubic meter (mg/M3) occurred between 1979 and 1982, but none thereafter. Comparison of the "whole state" panel with the other three panels clearly shows that high individual concentrations during the 1979-1982 period occurred in the Metro East area. The marked improvement in maximum and median Metro East airborne Pb concentrations after 1982 is likely to have been strongly influenced by the closure and subsequent cleanup of a secondary lead smelter in Granite City (Cooper and Frazier, 1983).

The national and state standard for Pb is a quarterly arithmetic mean of 1.5 mg/M3. No violations of this standard have occurred in Illinois since 1982.

Inspection of figure 3 indicates systematically higher individual Pb concentrations (open circles) in the Metro East area than in the Chicago area or the remainder of the state, especially during the 1979-1984 period. The statistical test for differences between regional means found only the Metro East Pb concentrations higher than the remainder of the state at the 5 percent level. Median Pb concentrations in the Metro East area were also higher than those in the Chicago area, but the differences were not quite significant at the 5 percent level.

As Pb concentrations have decreased, and violations of the standard have dropped to zero, the number of sites measuring airborne Pb concentrations in Illinois has also dropped sharply. Statewide, the number of sampling sites with valid annual means decreased from a high of 101 in 1980 to 17 in 1990. In the Chicago area the corresponding numbers are 60 and 8; in the Metro East area, 16 and 4; and in the remainder of the state, 26 and 5.

Nitrogen Dioxide (NO2). Concentrations of NO2 decreased substantially in Illinois during the 1980s. Table 2 shows decreasing trends for 1-hr and 24-hr maxima, as well as annual mean NO2 concentrations, both statewide and in the Chicago area. The trends range from 4.4 percent to -7.1 percent per year, and all are significant at the 1 percent level. There were not enough sampling sites in the other areas of the state to justify testing. Box plots for 1-hr maximum, 24-hr maximum, and annual mean NO2 are shown in figures 4-6, respectively. They illustrate the decreases, during the 1979-1990 period, in NO2 concentrations both statewide, and in the Chicago area, over all three averaging times. Although only the trend of the median has been tested statistically (table 2), it is apparent that concentrations associated with all the percentiles depicted by the box plots have decreased. The same is true for the highest individual values each year (plotted as open circles).

No national or state standards exist for 1-hr or 24-hr maximum NO2 concentrations. Numerous concentrations above the annual mean NO2 standard (0.053 ppm) occurred in the late 1970s, but none have occurred since 1980.

Almost all of the NO2 measurement sites are in the Chicago area, so it is impossible to look for differences in concentrations between geographic regions.

The number of sites measuring valid 1-hr NO2 concentrations has increased recently, both statewide and in the Chicago area, from a low of four sites in the early 1980s to about 15 sites in 1990. The number of sites with valid 24-hr and annual mean NO2 concentrations has decreased from 36-47 sites in the late 1970s to 10-15 sites in 1990.

Ozone (O3). All pollutants are affected to some extent by weather conditions. Poor dispersion conditions caused by low wind speeds or calm conditions can allow pollutant concentrations to build up in local areas near emission sources, and long- range transport can carry pollutants from their sources for deposit at distant locations. For O3, which is produced in the atmosphere by reactions between hydrocarbons and nitrogen oxides, however, weather conditions-especially temperature-affect how much of the pollutant is produced. Thus, it is desirable to remove the effect of weather conditions when looking for trends in O3. We first discuss O3 concentrations as observed, and then use a statistical regression approach to look for trends in O3 after the effect of temperature has been removed.

Table 2 indicates that no significant trend could be detected in the median of 1-hr maximum O3 concentrations in any geographic region of Illinois. The box plots of 1-hr maximum values by geographic area are shown in figure 7. Year-to-year variations are apparent, including higher concentrations statewide and in the Chicago area in 1983 and 1988, when abnormally high temperatures and little rain occurred during the summer ozone season, but overall the visual impression of no trend toward higher or lower concentrations agrees with table 2. The Chicago and Metro East areas clearly had higher median concentrations during the data period, compared to the remainder of the state, and these two urban centers also account for all the highest individual site values statewide (open circles).

The national and state standard for 1-hr maximum O3 concentration is 0.12 ppm. It is apparent that numerous violations of the standard occur most years in Illinois, although only two sites exceeded the standard (one time each) in 1990. Weather conditions over much of Illinois, and especially in the Chicago area, during the summer of 1990 were cooler, cloudier, rainier, and windier than normal (IEPA, 1991). These conditions undoubtedly contributed to the lower O3 concentrations observed.

The test for differences between regions showed that mean concentrations in the Chicago and Metro East areas were each significantly (5 percent level) greater than that in the remainder of the state.

A number of attempts to account for effects of atmospheric conditions on O3 concentrations have been published (e.g., Korsog and Wolff, 1991; Shively, 1991). Some of the past work has been quite sophisticated in searching for the best meteorological parameters to use in predicting O3 concentrations, and in the use of statistical methods to characterize the relationship. In contrast, the following attempt to account for the effects of meteorological conditions on O3 in Illinois is relatively simple. It uses only mean surface temperature during the O3 season (April- October) in a regression equation involving a constant, a temperature term, and a time term. The model is:

[O3] = Constant + á1 ú (Time) + á2 ú (Temperature) (1)

where [O3] is ozone concentration, and á1 and á2 are coefficients of time and temperature, respectively.

The procedure is to evaluate the coefficients of the time and temperature terms using observed O3 and temperature data, and to observe whether the coefficient of the time term is significantly different from zero. If it is, then there is a significant trend with time after the effect of temperature on O3 concentrations has been removed. The model was evaluated for each of the four geographical areas defined above. The O3 data were the medians of the annual distribution of 1-hr maximum O3 concentrations over all measurement sites in each geographical region. Corresponding April-October mean temperatures were obtained for each region from the Midwestern Climate Information System (Kunkel et al., 1990). Time was sequential year number, beginning with 1 in 1978. The regression analyses were carried out using the Multiple General Linear Hypothesis (MGLH) regression routine in the SYSTAT statistical software (Wilkinson, 1990). Results are shown in table 3.

The first thing to notice about the results is that the temperature coefficient is positive, indicating a positive correlation between O3 and temperature, as expected, in all four geographical regions. The temperature coefficient is significantly different from zero (5 percent level) over the whole state, and also in the "remainder" region, significant at the 10 percent level in the Metro East region, and just barely misses the 10 percent significance level (P = 0.1011) in the Chicago area.

The coefficient of the time term is negative, indicating a downward trend in concentrations, in all four regions. The downward trend, after accounting for the effects of surface temperature, is significant at the 5 percent level over the whole state and in the Chicago area, and at the 10 percent level in the other two regions. (Note: although not shown in table 2, a -1.9 percent per year downward linear trend in O3 significant at the 10 percent level was detected in the O3 concentrations as observed.)

A further indication of the importance of the time term is the improvement in the squared multiple correlation (R2 or the percent of the total variance explained by the regression) as the time term is added to the regression model. The model was evaluated separately without the time term to obtain R2 values with and without the time term in each region. Results are shown in table 4. There is a substantial improvement in the percent variance, explained with the addition of the time term in each region-from 26 to 58 percent for the whole state, from 5 to 46 percent in the Chicago area; from 16 to 42 percent in the Metro East area, and from 40 percent to 55 percent in the remainder of the state.

The results regarding ozone time trends may be summarized as follows. Without accounting for temperature effects, no trends were found at a 5 percent significance level, but a 1.7 percent per year downward trend in concentrations, significant at the 10 percent level, was detected in the Chicago area (not shown in table 3). After accounting roughly for temperature effects, downward trends were detected at the 5 percent level statewide and in the Chicago area, and downward trends significant at the 10 percent level were detected in the other two regions. For evaluating potential effects of ozone on agricultural crops, concentrations in the non-urban "remainder" region would be the most relevant.

Sulfur Dioxide (SO2). Table 2 shows downward trends significant at the 1 percent level for 3-hr maximum SO2 in the Chicago area (-3.5 percent per year), and for annual mean SO2 over the whole state (- 2.6 percent per year). The table also shows a downward trend significant at the 5 percent level for annual mean SO2 in the Chicago area (-5.2 percent per year). No trends significant at 5 percent or better were detected for any of the three types of observations in any region of the state; note, however, that the number of observing sites in the Metro East area was insufficient to estimate a trend.

Box plots for 3-hr maximum SO2 are shown in figure 8. Any trends are difficult to perceive by the naked eye, despite the highly significant trend detected in the Chicago area (table 2). Comparison of SO2 concentrations between regions shows that the highest 3-hr maximum SO2 values observed statewide occurred in areas other than the Chicago area. There is no national or state primary standard for 3-hr maximum SO2. The secondary national standard for 3-hr maximum SO2 is 0.5 ppm. Figure 8 shows that this standard has been exceeded on a few occasions during the 1978-1990 period. The test for differences between regions found concentrations in the Chicago area significantly lower than those in the rest of the state. The number of sampling sites with valid data for 3-hr maximum SO2 has remained very constant during the period.

Box plots for 24-hr maximum SO2 are shown in figure 9. No trends are apparent to the naked eye, and no significant trends were detected in table 2. Again, the Chicago area experienced somewhat lower median concentrations than the remainder of the state, which in this case includes the Metro East area. The difference was significant at better than the 1 percent level. The national and state primary standard for 24-hr maximum SO2, 0.14 ppm, has been exceeded a few times statewide in most years between 1978 and 1990; but only four of these occurred in the Chicago area. Numbers of sites with valid data fell by a factor of about 2 statewide over the 1978-1990 period. Most of this reduction occurred in the Chicago area, which had 47 sites with valid data in 1979, but only 11 sites in 1990. Most of this reduction is the result of closing sampling sites, rather than active sites not meeting completeness criteria.

Figure 10 shows box plots for annual mean SO2 concentrations. Decreases in median concentrations statewide are significant at the 1 percent level and in the Chicago area at the 5 percent level (table 2). Regional maximum concentrations (highest open circle) also appear to be on downward trends, at least for the "whole state" and "remainder" regions. There were no significant (5 percent) differences in concentration between geographic regions. Numbers of sites with valid data have also decreased over the 1978-1990 period, statewide by about 50 percent, and by about 70 percent in the Chicago area.

Particulate Mass (TSP, PM10). The state standard for particulate mass is written for total suspended particulate matter (TSP), while the national standard is for PM10, or particulate matter up to 10 mm in aerodynamic diameter.

Box plots for 24-hr maximum TSP concentrations are shown in figure 11. Median values have remained relatively constant over the 1978-1990 period with occasional anomaly years such as 1983, when spring dust storms in east central and northeastern Illinois (IEPA, 1984) caused numerous measurements > 500 mg/M3. No significant trends in median concentrations were detected (table 2) in any of the four geographic regions for the entire 1978-1990 period, but careful examination of the individual regional plots suggests a decreasing trend for the first half of the 1980s followed by an increasing trend in the later half, at least in some regions.

It is important to know whether these recent concentration increases reflect actual regional environmental conditions, or whether they are an artifact of the year-to-year changes in the measurement network. Numbers of sites with valid samples have dropped sharply in all regions in recent years, and if the sites removed from the network were systematically in cleaner areas (which might be expected), then average concentrations at the remaining sites would be higher. This question could be investigated by comparing records of concentrations at removed sites and at remaining sites.

The test for differences between regions found none (5 percent level), but some violations of the standard have occurred each year in all regions.

Trends in annual geometric mean TSP concentrations are shown in figure 12. No overall significant trends were detected (table 2), but the pattern of decreasing concentrations up to about 1985, followed by increases through 1989, is clear in all four regions. Again, however, the reality of the increase in the late-1980s is in doubt because the number of sites with valid data declined sharply during the 1986-1990 period. Comparison of concentrations between regions showed that concentrations in the Metro East area were significantly (1 percent level) higher than those in both the Chicago area and the remainder of the state. In addition, visual examination suggests that the highest individual site values occurred in the Metro East region.

The first published data for PM10 are for 1987 (IEPA, 1988), after the U.S. EPA changed its particulate matter standard from TSP to PM10, although some measurements were being made in anticipation of the new national standard as early as 1984 (IEPA, 1988). Because data were available for only four years, no evaluation of PM10 trends was included in table 2. However, box plots for 24-hr maximum and annual mean PM10 are shown in figures 13 and 14, respectively.

Figure 13 shows a number of violations of the 150 mg/M3 standard for 24-hr maximum PM10 in both the Chicago area and other areas of the state. The number of sampling sites increased in all three geographic areas shown in the figure during the 1987-1990 period.

Figure 14 also shows a few concentrations above the 50 mg/M3 standard statewide, none of which occurred in the Chicago area.

This concludes the presentation of data on the criteria pollutants. There are currently no standards for the remaining pollutants, which all occur as particles in the atmosphere, and are collected on high-volume filters.

Filter Analyses: Nitrate, Sulfate, and Metals

Through 1984, annual mean concentrations of nitrate, sulfate, and metals were derived from analyses of filters composited monthly. Thus, distributions of annual means are available from 1978-1990, but distributions of 24-hr maximum concentrations are available only after 1985, when measurements of individual 24-hr filters began.

Since all of these analyses depend on the collection of filter samples, it is appropriate to discuss changes in the number of valid samples that apply to all of the following analytes. Of course there are small individual differences between pollutants in the number of valid samples, but the major trends are determined primarily by the number of active sampling sites.

Statewide, the number of sampling sites collecting filter samples dropped from >100 in 1978 to <20 in 1990; the change from 1985 to 1990 was from about 50 sites to <20. In the Chicago area the drop in sites was from > 70 in 1978 to 35-40 in 1985 to <10 in 1990. In the Metro East area, there were > 15 sites up to 1981 and <5 after 1985. The numbers for the remainder of the state were > 25 sites up to 1981 and 4-6 thereafter.

Nitrate Ion (NO3-). Box plots for 24-hr maximum and annual mean NO3- concentrations are shown in figures 15 and 16, respectively. High year-to-year variability in the median (and other box plot percentiles) 24-hr maximum NO3- concentrations (figure 15) is apparent, particularly in the Metro East and "remainder" areas, where the plots are based on <10 sites.

For annual mean NO3-, the only significant trend (1 per-cent level) was a -2.1 percent per year decreasing trend in the "remainder" region (table 2). Figure 16 shows a relatively steady median concentration of about 5 mg/M3 statewide and in the Chicago area, and somewhat lower concentrations, with more year-to-year variability in the Metro East and remaining areas of the state. The comparison between regions found the Chicago area concentrations significantly (5 percent or better) higher than those in the Metro East area and the remainder of the state.

Both statewide and in the Chicago area, the variability within individual years, as measured by the interquartile (75th-25th percentile) range, or the height of the "box" in the box plot, appears to have decreased in recent years. Again, however, this occurred along with a steep decline in the number of sites with valid data, so the smaller variability may simply reflect the fact that measurements are occurring at fewer locations.

Sulfate Ion (SO42-). Box plots for 24-hr maximum and annual mean SO42- are shown in figures 17 and 18, respectively. Here also relatively high year-to-year variations in 24-hr maxima occurred, particularly in years or geographic areas with <10 sites.

Table 2 shows that declines in annual mean SO42- concentration significant at the 5 percent (whole state and Chicago area) or 2 percent (remainder of the state) levels occurred in all areas other than the Metro East area. The significant linear trends were all between -1.3 and -1.6 percent per year.

Concentrations were significantly (1 percent level) higher in the Metro East area than in the Chicago and "remainder" regions of the state.

Arsenic (As). Box plots for 24-hr maximum and annual mean As are shown in figures 19 and 20, respectively. Except for the Metro East area, where median 24-hr maximum concentrations >0.01 mg/M3 occurred in four of the six years of record, the median 24- hr maximum was <0.01 mg/M3 statewide. One value > 0.5 mg/M3 was observed in the Metro East area in 1990.

Differences between geographical regions are the most interesting feature of these plots. Comparison of the plots for the various regions shows clearly that the Metro East area accounts for the highest measured values statewide; this is true for both maximum 24-hr concentrations as well as mean annual concentrations. The statistical test for differences between regions was not run on the 24-hr maximum As data because of the relatively few years of data. For annual mean As, the test found that concentrations in the Metro East area were higher than those in the Chicago and "remainder" areas at the 0.1 percent significance level.

Table 2 shows declines in As concentrations of -8.4, -11.1, and -9.5 percent per year in the whole state, the Chicago area, and the "remainder" region, respectively. All are significant at the 1 percent level.

Cadmium (Cd). Box plots for 24-hr maximum and annual mean Cd are shown in figures 21 and 22, respectively. Again, differences between geographical regions are clear, and again the Metro East area accounts for most of the highest values that were observed statewide over both averaging times. The test for regional differences in annual mean concentration confirmed that the highest concentrations occurred in the Metro East area at a significance level of 0.1 per-cent. Table 2 shows that the decline of -12.7 percent per year in annual mean Cd in the "remainder" region was significant at the 2 percent level.

Chromium (Cr). Figures 23 and 24 show box plots of 24-hr maximum and annual mean Cr, respectively. Note that Cr was not measured on filter samples until 1985. The number of sites in the Metro East area was insufficient for a separate plot, so the Metro East sites are included in the "remainder" region. The marked differences between regions that were apparent for As and Cd are smaller or absent here. Rather than the Metro East area having the highest concentrations, Cr appears to be somewhat more prominent in the Chicago area, although the differences were not statistically (5 percent) significant. No significant trends in the median annual mean Cr were detected (table 2).

Iron (Fe). Box plots for 24-hr maximum and annual mean Fe are shown in figures 25 and 26, respectively. For both averaging periods, it is again clear from the figures that the Metro East area accounts for the highest values statewide. The test for differences between regions confirmed that the highest annual mean Fe concentrations occurred in the Metro East area (0.1 percent level). Moreover, the trend test (table 2) detected an upward trend (6.0 percent per year) in the Metro East area, significant at the 2 percent level. While it is true that the two area sampling sites with the lowest annual mean concentrations were closed during the 1985-1990 period (one after 1985, and the other after 1988), this is not the cause of the increasing trend in concentrations. Concentrations at the remaining sites also increased during these years.

Manganese (Mn). Figures 27 and 28 show box plots for 24-hr maximum and annual mean Mn, respectively. Both the Chicago area and the Metro East area contributed to the highest values over both averaging periods observed statewide, but the Metro East area experienced noticeably higher median concentrations (figures 27 and 28). This was confirmed for annual means by the test for differences between regions, which found higher concentrations in the Metro East area compared to the Chicago and "remainder" areas (0.1 percent level). As in the case of Fe, the trend test (table 2) detected an increasing trend (10 percent per year), significant at the 5 percent level, in annual mean Mn in the Metro East area. Also as in the case of Fe, the closing of sampling sites could have had only a minor effect on the increasing Mn concentrations over the 1985-1990 period. Nickel (Ni). Box plots for 24-hr maximum and annual mean Ni are shown in figures 29 and 30, respectively. Ni was not measured on filter samples until 1984. Since the number of sites in the Metro East area was insufficient for a separate plot, the Metro East sites are included in the "remainder" region. For Ni, it appears that the Chicago area contributed most of the highest 24-hr maximum concentrations (figure 29). Comparing annual means between regions, it appears that the Chicago area has recently experienced somewhat higher median concentrations, but the statistical test found no significant (5 percent level) differences. Apparently, both regions contribute more or less equally to the highest annual means. The trend test (table 2) found no significant trends in the median annual mean Ni.Metal Enrichments, Relative to Soil. Enrichment factors are a common and convenient method of distinguishing between natural and anthropogenic sources of metals in the atmosphere. For many metals the typical natural source examined is crustal or soil aerosols. Enrichment, E, for element X is expressed as a ratio of ratios: E = (X/RE)air/(X/RE)soil (2) where RE is a reference element for which the soil or earth's crust is the dominant source in the atmosphere. Enrichments are usually evaluated for measured concentrations of X in air, using abundances (mass fractions) of X and RE in soils from compilations of global mean soil or crustal composition. Figure 31 shows element enrichments based on the median statewide annual mean concentrations of As, Cd, Cr, Pb, Mn, and Ni, using Fe as the reference element, for 1978-1990. If the ratio of the subject element to Fe in air is the same as the ratio of the same elements in soil, then the enrichment is 1.0, and the element X is assumed to come only from the soil. In practice, because of variations in soil composition, elements with enrichments between 1 and 10 are usually assumed to have primarily soil sources. Values of E >10 are assumed to indicate anthropogenic sources. Using these criteria, figure 31 shows that Pb and Cd, with enrichments between 100 and 3000, have primarily anthropogenic sources in Illinois. Note that the enrichment of Pb has dropped from 1000-2000 in the early 1980s to 100-200 in recent years as Pb has been removed from automotive fuels. Nevertheless, the concentration of Pb currently in the atmosphere is far in excess of what would be there naturally from soil wind erosion. Sources such as Pb smelters and fly ash from coal burning may now be the major sources of Pb in the atmosphere, although Pb recycled from previously deposited auto exhaust may also be important. Cd also has a high enrichment factor of about 1000, which has remained relatively constant during the 1980s. The absence of a Cd data point in 1981 is the result of a one-year hiatus in Cd measurements. Those in 1983 and 1989 were caused by median Cd concentrations of 0.000 mg/M3. Arsenic enrichments are mostly >10 also, suggesting a mostly anthropogenic source. They appear to be decreasing somewhat, in agreement with the decreasing trends in concentration seen in table 2. Mn enrichments are firmly in the 1-10 range generally ascribed to natural sources, with very little year-to-year variation. The majority of Mn in the Illinois atmosphere may well be natural. Mn sources also tend to be Fe sources, however, so the extent to which the reference element Fe is nonnatural, may also hold true for Mn. For example, anthropogenic sources are very likely to be the cause of the significant increase (at the 5 percent level, or better) in concentrations of both Fe and Mn during the 1980s in the Metro East area. The year-to-year variations in Cr and Ni enrichments during the 1985-1990 period are similar to each other, but unlike those of any of the other metals in figure 31. The enrichments are mostly in the 1-10 range, suggesting mainly natural sources for these metals, but the sometimes large year-to-year variations suggest that concentrations of both of these metals are enhanced by anthropogenic sources at least during some years.

SPATIAL DISTRIBUTIONS OF POLLUTANT CONCENTRATIONS

It was a very early goal of this summary of Illinois air quality to draw spatial contours of pollutant concentrations on a statewide scale. Examination of the data quickly dashed that early hope, however, as it became clear that sampling sites were concentrated in the two major industrial and population centers, the Chicago and the Metro East areas. The relatively few sampling sites not in these major centers were clustered in smaller cities and the environs of troublesome sources. Data are not available from sparsely populated regions with no major sources. For an agency such as the IEPA with the mission of documenting problems and enforcing standards, this sampling strategy is obvious. It does not provide data needed to document statewide air quality, however. For this report, we propose that the data are adequate to show spatial variations in that portion of the state with the highest concentration of sampling sites, the Chicago area. To provide a sense of the temporal changes in Chicago-area spatial patterns, contours have been drawn for 1980, 1985, and 1990, and are presented below. The rapid drop in the number of sampling sites, even in the Chicago area, makes the 1990 contours very problematic. The 1990 contours are meant to suggest only the most rudimentary spatial patterns. For most pollutants the data are not adequate to do more. The data and associated discussions for the several pollutants appear in the same order as those of the temporal trends above. Criteria Pollutants CO. Figures 32 and 33 show spatial distributions of 1-hr maximum and 8-hr maximum CO concentrations in the Chicago area in 1980, 1985, and 1990. The low density of sampling sites, and the changes in site locations between years preclude identification of persistent locations of high or low concentrations. In 1980 one sampling site near the lake shore had a 1-hr maximum >20 ppm. In 1985 three sites had 1- hr maxima between 10 and 20 ppm, but there were no measurements >20 ppm. In 1990, only one site recorded a 1-hr maximum >10 ppm. This general decline in concentrations agrees with the significant overall decline in Chicago-area concentrations detected by the trend test (table 2). No violations of the 35 ppm national and state standard for 1- hr maximum CO were observed. Figure 33 shows spatial patterns of 8-hr maximum CO concentrations. Again, temporal changes in the sampling network limit discussion mostly to temporal changes in concentrations. In 1980, one lakeshore site exceeded the national and state standard for 8-hr maxi-mum CO (9 ppm), but no violations were observed in the region in 1985 or 1990. A general decline in concentrations is apparent from 1980 to 1985 to 1990 that agrees with the significant decline detected in the Chicago area from the full 1978-1990 dataset (table 2).Pb. Spatial distributions of annual mean Pb concentrations in the Chicago area for 1980, 1985, and 1990 are shown in figure 34. Persistent concentration maxima occurred in the central city and the southeast Chicago industrial area; however, the well-known and very significant decline (table 2) in atmospheric Pb concentrations during the 1980s is evident here as well. The value of the highest contour line declined from 0.6 mg/M3 in 1980 to 0.2 mg/M3 in 1985 and 0.1 mg/M3 in 1990. Pb standards are written in terms of mean concentrations over calendar quarters, which cannot be inferred from the data shown here. See the earlier discussion related to temporal trends. NO2. Spatial patterns of 24-hr maximum and annual mean NO2 in the Chicago area are shown in figures 35 and 36, respectively. In 1980 the highest 24-hr maxima occurred in suburban northwestern (DesPlaines) and southern (Flossmoor) Cook County, with single sites in each area experiencing one or more days with mean concentrations >0.14 ppm. By 1985, the two sites with the highest values in 1980 were no longer active. The data suggest an overall decline in maximum 24-hr concentrations, with the highest value (0.094 ppm) occurring in the city center. The 1990 data show a further overall decline in maximum 24-hr concentrations, with the maximum (0.062 ppm) occurring in a cluster of sampling sites near O'Hare Airport. There are no national or state standards for 24-hr maximum NO2. Figure 36 shows Chicago-area contours of annual mean NO2. Shaded areas in the northwest suburbs (DesPlaines) and the central city exceeded the national and state primary standard (0.053 ppm) in 1980. Although no violations of the standard were observed in 1985 and 1990, the available data suggest that the highest annual mean concentrations still occurred in the central city area near Lake Michigan. The apparent decreasing concentrations agree with the significant trend in Chicago area annual mean NO2 discussed earlier (table 2).O3. Figure 37 shows contours of maximum 1-hr O3 concentrations in the Chicago area for 1980, 1985, and 1990. Most of the Chicago area was in violation of the 0.12 ppm standard in 1980, with a conspicuous area of lower values across the middle of Cook County and the city of Chicago. In 1985 the violations of the standard were confined to the northeastern and southeastern portions of Cook County. No violations were observed in 1990; however, there is evidence that this may have occurred because summer weather conditions were relatively unfavorable for O3 formation, as discussed earlier. Table 2 indicates no significant temporal trend in O3 as observed; however, as shown earlier, when surface temperature is accounted for, there is a significant trend (2 percent level) toward decreasing concentrations in the Chicago area over the 1978-1990 period. SO2. Figures 38-40 show Chicago-area spatial contours for SO2: 3- hr maximum, 24-hr maximum, and annual mean concentrations, respectively. The contours drawn from the available data are very difficult to interpret. Overall, 3-. A summary of the test where maximum concentrations were primarily >0.10 ppm in 1980 and 1985, and primarily <0.10 in 1990. This apparent trend toward decreasing concentrations agrees with the significant trend for the Chicago area detected earlier (table 2). A single high value (0.446 ppm) was observed in the western suburbs (Bedford Park) in 1985. There are no national or state standards for 3-hr maximum SO2. Chicago-area spatial contours for 24-hr SO2 are shown in figure 39. One site in south-suburban Cook County (Blue Island) exceeded the national and state standard (0.14 ppm) in 1980. Aside from this extreme, concentrations at most locations were <0.05 ppm. The available data for 1985 and 1990 show maxima in the western suburbs (Bedford Park) in both years, but again most sites experienced 24-hr maximum concentrations <0.05 ppm. There were no violations of the standard in 1985 or 1990. Table 2 shows no significant temporal trends in 24-hr maximum SO2 in the Chicago area.Chicago-area spatial contours for annual mean SO2 concentrations are shown in figure 40. The figure shows no strong spatial patterns in any of the t> > three years, but very different patterns of minimal high and low concentrations over the three years. There were no violations of the national and state standard (0.03 ppm) for the annual mean in any of the three years. TSP. Spatial patterns of 24-hr maximum and annual mean TSP in the Chicago area are shown in figures 41 and 42, respectively. The spatial pattern for 1980 shows areas of high concentrations, including violations of the 260 mg/M3 standard, in (1) the industrial area of southeastern Chicago, (2) the western suburbs of Cook County (River Forest) and eastern dupe County (Elmhurst), (3) northwestern Will County (Romeoville, Joliet, and Rockdale), and (4) southwestern Will County (Braidwood). In 1985 there were widespread violations of the standard, over almost all of Cook County, the southern half of dupe County, and southwestern Will County On the other hand, there were no violations in the Joliet area. The widespread extent of the violations of the 24-hr standard suggests that a regional weather event might have been responsible, and indeed, a dust storm that occurred on May 31, 1985, was the cause of most of the violations (IEPA, 1986). In 1990 the much-reduced sampling network detected only one violation of the standard at Rockdale in Will County. Figure 42 shows the spatial patterns for annual geometric mean TSP. In 1980 maximum concentrations above the national standard (75 mg/M3) occurred in a broad band from the central city area of Chicago to suburban Summit, in the southeast Chicago industrial region, in northeast Dupe County (Bensenville), and in the Joliet area of Will County In 1985, only two sites in the Chicago area violated the standard, one in southeast Chicago and one in west suburban McCook. In 1990 only one site, in downtown Chicago, was in violation; however, the number of sampling sites was down considerably.

Filter Samples: Nitrate, Sulfate, and Metals. There are no standards for the remaining pollutants. NO3-. Spatial patterns of 24-hr maximum and annual mean NO3- in the Chicago area are shown in figures 43 and 44, respectively. Data on 24-hr maximum NO3- are not available before 1985. Maximum 24-hr NO3- concentrations are relatively uniform over the area in both 1985 and 1990, varying by a factor of only about 2 from minimum to maximum. Thus, the importance of the locations of relative high and low concentrations is difficult to assess. Figure 44 shows spatial patterns of mean annual NO3- in the Chicago area. Again, the values are spatially very uniform. It is interesting to note that in 1980 the Cook County concentrations were generally <5 mg/M3 while those in DuPage and Will Counties were > 5 mg/M3; in 1985 the pattern was reversed, although that assessment is based on only four sites outside of Cook County There were too few sites in 1990 to justify any comment on spatial variations. SO42-. Spatial patterns of 24-hr maximum and annual mean SO42- in the Chicago area are shown in figures 45 and 46, respectively. Interpretation of these patterns is limited by the same factors that affected the NO3- data: lack of data for 1980 and a large disparity between the number of sampling sites in 1985 and 1990. Based on only four sampling sites, the data provide a hint that maximum 24-hr concentrations >40 mg/M3 affected more of the Cook County area in 1990 than in 1985. Further, the southeast Chicago industrial area had maximum concentrations >40 mg/M3 in both years. Figure 46 shows Chicago-area spatial patterns of annual mean SO42-. The areas where mean annual SO42- exceeded 15 mg/M3 were scattered throughout Cook County in 1985, but were reduced to one site in 1985, and none at all in 1990. Fe. Spatial distributions of 24-hr maximum and annual mean Fe concentrations in the Chicago area for 1985 and 1990 are shown in figures 47 and 48. In 1985, maximum 24-hr concentrations varied between 1 and 4 mg/M3 over most of the three-county Chicago area, but there was a small area where concentrations exceeded 20 mg/M3 in the southeast Chicago industrial area. In 1990 the typical range was still 1-4 mg/M3, but the high concentrations in southeast Chicago were down considerably. Figure 48 shows spatial patterns of mean annual Fe concentrations in the Chicago area. The main feature of the 1980 pattern was a broad area of concentrations >1.0 mg/M3 in central and southeastern Cook County Note, however, the lack of 1980 measurements in the southeast Chicago industrial region, where mean values >3 mg/M3 were observed in 1985. Most of the few sampling sites remaining in 1990 reported mean Fe concentrations >1.0 mg/M3, but the southeast Chicago hot spot was gone. Mn. Spatial patterns of 24-hr maximum and annual mean NO3- in the Chicago area are shown in figures 49 and 50, respectively. Notice the similarities between the spatial patterns of Mn (figures 49 and 50) and Fe (figures 47 and 48), especially in 1985. This is not surprising since baseline concentrations are probably the result of wind erosion of soil dust, while local peak concentrations are often associated with industrial activity, especially steel- making. Peak 24-hr maximum concentrations of both Mn and Fe in southeast Chicago decreased by a factor of 4 or 5 from 1985 to 1990. Because of very limited numbers of sampling sites in the Chicago area through 1985, spatial data for the remaining four metals, As, Cd, Cr, and Ni, is limited to 1990 annual means, all presented in figure 51. As. Figure 51 shows annual mean concentrations of As at eight sites in the Chicago area. Mean concentrations at six of the eight sites were 0.001 mg/M3, and the other two were 0.000 mg/M3. The figure presents only the data. No contours are justified. Cd. The 1990 spatial pattern of Cd in the Chicago area is also shown in figure 51. Annual mean concentrations were 0.001 or 0.002 mg/M3 at six of the eight sites, with a maximum value (0.008 mg/M3) in southeast Chicago. Cr. The 1990 spatial pattern of Cr in the Chicago area is also shown in figure 51. Annual means ranged from 0.000 to 0.015 mg/M3. The maximum concentrations occurred in southeast Chicago and the western suburb of Maywood. Ni. The Chicago-area pattern of mean annual Ni concentrations for 1990 are also shown in figure 51. There was a broad concentration maximum (>0.005 mg/M3) over central and southern Cook County, with the highest value at Maywood.

TABULATIONS OF LITERATURE DATA ON OCCASIONAL MEASUREMENTS

Some pollutants, including some considered toxic, have not been measured routinely in Illinois for reasons that usually involve the cost of analysis, the ability to detect the extremely low concentrations present in the atmosphere, or both. Table 5 summarizes some recent measurements of volatile organic compounds in Illinois. Measurements in table 5 were made in Chicago, in the Metro East area, and at a rural agricultural site near Champaign between 1986 and 1990. Sweet and Vermette (1992) found that the concentrations they measured in Illinois urban areas were quite similar to those observed in other urban areas of the United States. Further, the concentrations they observed in an industrial area were quite similar to those seen in other cities without significant industry, which suggests that the major types of sources were wide-spread small sources, especially related to automotive emission and fuels, rather than large point sources. The available data are not adequate to address the issue of temporal trends, except that there are a few substantial differences in mean concentrations between years. With regard to spatial variations, there are compounds for which rural and urban concentrations are similar or different, and some differences in concentration between urban areas. As is true for many atmospheric parameters, concentrations of many volatile organic compounds are highly variable both spatially and temporally. These variations depend on sample duration, proximity of sampling to sources, variations in source strength, and especially on weather conditions. Variability between years is seen by comparing the various measured mean concentrations in Chicago. In general, the results of McAlister et al. (1989, 1991) for 1988 and 1990 agree quite well with those of Sweet and Vermette (1992) measured over the 1986-1990 period. However, there are a few exceptions to this rule, notably carbon tetrachloride and tetrachloroethylene in 1990, and toluene and metaplus paraxylene in 1988. The measurements of Wadden et al. (1992) were made in a different Chicago location, near the city center. There is good agreement between the Wadden et al. measurements and those already mentioned for some compounds, such as benzene, but poor agreement for others, such as chloroform and ethylbenzene. In most cases the Wadden et al. values are greater than the others, probably because of spatial variations within the city. Variability between concentrations of organic compounds between urban areas occurs for some compounds, and not for others. For benzene, chlorobenzene, ethyl benzene, trichloroethylene, and metaplus paraxylene, mean concentrations show differences ranging from factors of 2 to 10. Note, however, that the variability of the respective concentrations is very high, judging from the standard deviation, so that the differences may not be statistically significant. It is noteworthy that the Metro East concentration was higher for each of these compounds. Urban/rural variability, or to be more precise, the lack of differences in concentration between urban and rural locations, is seen for a few compounds, notably carbon tetrachloride and chloroform (Sweet and Vermette, 1992). Such uniform concentrations are evidence of a well-mixed contaminant with a long atmospheric residence time and few current sources. Carbon tetrachloride is an example of a compound no longer used commercially, so its current sources are very small. However, its residence time in the atmosphere is long, so over time it has become well mixed in the atmosphere, and its concentrations are relatively uniform regardless of whether the sampling site is in a rural or urban area (Sweet and Vermette, 1992).

DISCUSSION

Trends over Time: Increasing, Decreasing, and Level Concentrations Criteria Pollutants. Six pollutant types are included among the criteria pollutants: CO, Pb, NO2, O3, SO2, and particulate matter. Some of these are measured over multiple averaging times, but there is not necessarily a standard for every averaging time. For example, NO2 is measured over averaging times of 1 hour, 24 hours, and the calendar year, but there are state and national primary standards only for the annual mean. Particulate matter is measured as TSP, for which there are only state standards (for both annual geometric mean and 24- hr maximum concentrations), and as PM10, for which there are only national standards (annual arithmetic mean and 24-hr maximum). Thus, for the six pollutant types, measurements have been made for 14 separate pollutant-averaging time combinations, 11 of which have state or national primary standards. The PM10 measurements have only been made for a few years, not enough to make a trend test meaningful, so of the 14 combinations, only 12 (at the most) have been tested for time trends within individual geographic regions. In some regions there were not enough sites to carry out trend tests. The results of the trend tests within regions are summarized in table 2.For the state as a whole, 12 pollutant/averaging time datasets were tested for time trend. Seven of these (table 2) showed trends toward decreasing concentrations significant at the 1 percent or 2 percent level, four had no significant trend (5 percent) (including O3, before accounting for temperature effects), and none showed increasing trends. After accounting for temperature effects, O3 showed a decreasing trend (2 percent level, table 3), rather than no trend. The same 12 kinds of datasets were tested for trend in the Chicago area. Eight of these (table 2) showed decreasing trends at the 5 percent level (all but one of these were at the 1 or 2 percent level). Four showed no significant trend at the 5 percent level, and none showed an increasing trend. After accounting for temperature, O3 showed a decreasing trend (2 percent level, table 3), rather than no trend. In the Metro East area, only four pollutant/averaging time datasets were measured at enough sites to warrant testing for time trend. Only Pb showed a decreasing trend (1 percent), while three other datasets showed no significant trend (5 percent). After accounting for temperature, O3 showed a decreasing trend at the 6 percent level, but not at the 5 percent level. No significant (5 percent) increasing trends were found for criteria pollutants. Nine datasets were tested for time trend in the "remainder" region, four of which included a few sites from the Metro East region when there were insufficient sites for separate tests there. Three of these datasets showed decreasing trends at the 5 percent level (all but one at the 1 percent level), and five datasets showed no significant (5 percent level) trends. After accounting for temperature, the decreasing trend of O3 with time was significant only at the 10 percent level. Again, there were no significant increasing trends. Aside from Pb, the largest decreasing linear trends were found for annual mean NO2, at -7.1 percent per year statewide and -6.8 percent per year in the Chicago area. The greatest magnitudes of decreasing linear trends were found for Pb: - 12.6 percent per year in the Metro East area, and -20 to -25 percent per year statewide and in the other geographic areas. Noncriteria Pollutants. Only trends in annual mean concentrations of the noncriteria pollutants measured-NO3-, SO42-, and the six metals (As, Cd, Cr, Fe, Mn, and Ni)-were tested for trend over time. Currently, 24-hr average metal concentrations are also being measured, but their record is too short to permit a meaningful test of time trend. The eight noncriteria pollutants were all tested for trend in all four regions, except for the Metro East region, where the data were not adequate to test Cr and Ni. Of the 30 trend tests, 20 showed no significant (5 percent) trend. Over the state as a whole, and in the Chicago area, two species showed significant decreases: SO42- at the 5 percent level and As at the 1 percent level. In the Metro East area, Fe and Mn showed significant increases (the only increasing trends found in this study): Fe at the 2 per-cent level, and Mn at the 5 percent level. In the "remainder" region, SO42- (2 percent), As (1 percent), Cd (2 percent), and Mn (5 percent) showed significant decreases. Spatial Trends: Illinois Hot Spots Comparison of Geographic Regions. Comparison of median and maximum pollutant concentrations within geographic regions, from the yearly box plots in figures 1-30, indicates which geographic areas of the state experienced the highest concentrations of air pollutants. Chicago generally had higher median regional values of annual mean NO3-, and of annual mean and 24-hr Cr. It also experienced higher median annual mean Ni and the highest individual 24-hr Ni concentrations. These Cr and Ni results agree with the observations of Sweet and Gatz (1988), based on measurements of fine and coarse particles with dichotomous samplers. On the other hand, the Chicago area generally experienced lower concentrations of 3-hr and 24-hr SO2 than the rest of the state. The Metro East area generally experienced higher concentrations than the rest of the state for annual mean Pb, annual mean TSP, and both 24-hr and annual mean As, Cd, Fe, and Mn. Again, the As and Cd results agree with those of Sweet and Gatz (1988), but for Pb and Mn, Sweet and Gatz found higher concentrations in Chicago. The Chicago and Metro East areas experienced higher concentrations than the rest of the state for 1-hr maximum O3 and annual mean SO2. Chicago Area Concentration Contours. Based on the contour plots in figures 32-51, statements about preferred areas of relatively high pollutant concentrations should be regarded as best guesses, based on available data. But these statements are uncertain as to magnitude of pollutant concentrations as well as to precise location. The available data are not adequate for highly certain conclusions because: 1) sampling sites are distributed nonuniformly over the area; 2) the plotted data represent only three years (1980, 1985, and 1990); and 3) numbers of sampling sites, and their locations, have changed from one plotted year to the next, and they have recently dwindled to the extent that the meaning of contours is questionable. Only one locale in the Chicago area stands out for being the location of high concentrations for multiple pollutants: the industrial area of southeast Chicago around Lake Calumet. This area has persistent, relatively high concentrations of some metals, in terms of both 24-hr maxima and annual means. The metals with highest 24-hr maxima in the Lake Calumet area were Fe and Mn. There was a persistent peak in 24-hr SO42- in this area as well. Metals with local peaks in annual mean concentrations include Pb, and possibly Cd, Cr, and Mn, although the evidence for the latter three is somewhat weaker. Other locations in the Chicago area appeared as areas of persistent high concentrations of only one or two pollutants. One-hour maximum CO was persistently high along the Lake Michigan shoreline north from downtown Chicago. Annual mean NO2 concentrations were also persistently high near the Chicago city center (the Loop). The plotted O3 data (figure 37) suggest a possible persistent area of high 1-hr maximum concentrations in northeast Cook County. There is also some weak evidence for relatively high 1-hr maxima of O3 and annual mean NO2. Southwest suburban Bedford Park had persistently high concentrations of 3-hr maximum, 24-hr maximum, and annual mean SO2. Finally, various locations in Will County experienced persistently high 24-hr maximum TSP concentrations, perhaps related to wind erosion of agricultural soils.

CONCLUSIONS

Temporal Trends Criteria Pollutants. Of all the criteria pollutants tested for time trend in any geographic region of the state, results indicated only decreasing trends or no significant (5 percent) trends. No increasing trends in criteria pollutants were detected. For the state as a whole, seven of 12 pollutant/averaging time datasets tested for time trend showed significant trends (5 percent level or better) toward decreasing concentrations over the 1978-1990 data period. Pb had the largest decreasing mean linear trend statewide, -20.5 percent per year. After accounting for temperature effects, O3 showed a significant decreasing trend (2 percent level), rather than no Trend. In the Chicago area, eight of the 12 datasets showed decreasing trends at the 5 percent level (all but one of these were at the 1 or 2 percent level). Again, Pb had the largest trend, -21.2 percent per year. After accounting for temperature, O3 showed a decreasing trend (2 percent level, table 3), rather than no trend, in the Chicago area. In the Metro East area, only four pollutant/averaging time datasets were measured at enough sites to warrant testing for time trend. Only Pb showed a significant (5 percent) trend, a trend toward decreasing concentrations. The Pb trend was sizable, -12.6 percent per year, but much smaller than the Pb trend observed in all other areas of the state. After accounting for temperature, O3 showed a decreasing trend at the 6 percent level, but not the 5 percent level. Three of nine datasets tested for time trend in the "remainder" region showed decreasing trends at the 5 percent level or better. Again, Pb had the largest trend, -24.8 percent per year. Noncriteria Pollutants. The eight noncriteria pollutants were all tested for trend in all four regions, except for the Metro East region, where the data were not adequate to test Cr and Ni. Of the 30 trend tests, 20 showed no significant (5 percent) trend. Over the state as a whole, and in the Chicago area, two species showed significant decreases-SO42- and As. In the Metro East area, Fe and Mn showed significant increases (the only increasing trends found in this study). In the "remainder" region, SO42-, As, Cd, and Mn showed significant (5 percent or better) decreases. Spatial Trends Comparison of median and maximum pollutant concentrations within geographic regions, from the yearly box plots in figures 1-30, indicates which geographic areas of the state experienced the highest concentrations of air pollutants. Chicago generally had higher median regional values of annual mean NO3-, and of annual mean and 24-hr Cr. It also experienced higher median annual mean Ni and the highest individual 24-hr Ni concentrations. On the other hand, the Chicago area generally experienced lower concentrations of 3-hr and 24-hr SO2 than the rest of the State. The Metro East area generally experienced higher concentrations than the rest of the state for annual mean Pb, annual mean TSP, and both 24-hr and annual mean As, Cd, Fe, and Mn. The Chicago and Metro East areas experienced higher concentrations than the rest of the state for 1-hr maximum O3 and annual mean SO2. Within the Chicago area, only one location stands out for its high concentrations of multiple pollutants. This is the industrial area of southeast Chicago around Lake Calumet. This area has persistent, relatively high annual mean or 24-hr concentrations, or both, of SO42-, Fe, Mn, and Pb, and possibly Cd, Cr, and Mn. The evidence for the latter three is somewhat weaker than for the others, however. Other locations in the Chicago area appeared as areas of persistent high concentrations of only one or two pollutants.

ACKNOWLEDGMENTS

Bob Swinford and Dave Kolaz, of the Illinois EPA, were very helpful in providing data, answering questions about the data, and discussing numerous issues. Sherman Bauer entered most of the data into computer files. Peter Scheff and Clyde Sweet provided data on concentrations of organic pollutants. Bob Sinclair provided guidance for conversion of the data to GIS files. Tom Heavisides supplied information on data sources.

REFERENCES

Appel, B.R., Y. Tokiwa, M.Haik, and E.L. Kothny. 1984. Artifact particulate sulfate and nitrate formation on filter media. Atmospheric Environment, 18(2):409-416.Cooper, J.A., and C.A. Frazier. 1983. Source apportionment of TSP and lead in Granite City, Illinois, using chemical mass balance receptor model methods. Final Report, Vols I and II, prepared for the Illinois Environmental Protection Agency by NEA, Inc. Illinois EPA, Springfield, IL.Illinois Environmental Protection Agency. 1979. Annual Air Quality Report 1978. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.

Illinois Environmental Protection Agency. 1980. Annual Air Quality Report 1979. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1981. 1980 Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1982. 1981 Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1983. 1982 Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1984. Illinois Annual Air Quality Report 1983. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1985. Illinois Annual Air Quality Report 1984. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1986. Illinois Annual Air Quality Report 1985. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1987. Illinois Annual Air Quality Report 1986. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1988. Annual Air Quality Report 1987. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1989. Illinois 1988 Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706.Illinois Environmental Protection Agency. 1990. Illinois 1989 Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706. Illinois Environmental Protection Agency. 1991. 1990 Illinois Annual Air Quality Report. IEPA, Division of Air Pollution Control, Air Monitoring Section, 2200 Churchill Road, Springfield, IL 62706. Korsog, P.E., and G.T. Wolff. 1991. An examination of urban ozone trends in the northeastern U.S. (1973-1983) using a robust statistical method. Atmospheric Environment, 25B(1):47-57. Kunkel, K.E., S.A. Changnon, C.G. Lonnquist, and J.R. Angel. 1990. A real- time climate information system for the midwestern United States. Bull. Amer. Meteorol. Soc., 71(11): 1601- 1609. McAlister, R.A., W.H. Moore, J. Rice, D.-P. Dayton, R.F. Jongleux, P.L. O'Hara, R.G. Merrill, and J.T. Bursey. 1989. 1988 nonmethane organic compound monitoring program, Vol. II: urban air toxics monitoring program. Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle Park, NC 27711, EPA-450/4-89-005. McAlister, R.A., E. Bowles, J. DeGarno, J. Rice, R.F. Jongleux, R.G. Merrill, and J.T. Bursey. 1991. 1990 urban air toxics monitoring program. Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle Park, NC 27711, EPA-450/4-91-024. Science Advisory Board, U.S. Environmental Protection Agency. 1990. Reducing Risk: Setting Priorities and Environmental Strategies for Environmental Protection. The Report of the Science Advisory Board, Relative Risk Reduction Strategies Committee. U.S. EPA, Science Advisory Board, SAB-EC-90- 021. Shively, T.S. 1991. An analysis of the trend in ground- level ozone using non-homogeneous Poisson processes. Atmospheric Environment, 25B(3):387-395.Snedecor, G.W., and W.G. Cochran. 1980. Statistical Methods, 7th edition, The Iowa State University Press, Ames, Iowa, p. 192.Sweet, C.W., and S.J. Vermette. 1992. Toxic volatile organic compounds in urban air in Illinois. Environ. Sci. & Technol., 26(1), pp. 165-173.Wadden, R.A., P.A. Scheff, J. Lin, H. Lee, C.B. Keil, J. Graf-Teterycz, D. Kenski, J.-Y. Yeng, M. Javor, N. Khalili, and T. Holsen. 1992. Evaluation of two-phase air pollution data for receptor modeling. Paper 92-104.05, 85th Annual Meeting of the A&WMA, Air & Waste Management Association, Pittsburgh, PA.Wilkinson, L. 1990. SYSTAT: The System for Statistics, SYSTAT, Inc., Evanston, IL.

Figure Captions Figure 1. Trends in 1-hr maximum CO concentrations in Illinois. Figure 2. Trends in 8-hr maximum CO concentrations in Illinois. Figure 3. Trends in annual mean Pb concentrations in Illinois. Figure 4. Trends in 1-hr maximum NO2 concentrations in Illinois. Figure 5. Trends in 24-hr maximum NO2 concentrations in Illinois. Figure 6. Trends in annual mean NO2 concentrations in Illinois. Figure 7. Trends in 1-hr maximum O3 concentrations in Illinois. Figure 8. Trends in 3-hr maximum SO2 concentrations in Illinois. Figure 9. Trends in 24-hr maximum SO2 concentrations in Illinois. Figure 10. Trends in annual mean SO2 concentrations in Illinois. Figure 11. Trends in 24-hr maximum TSP concentrations in Illinois. Figure 12. Trends in annual geometric mean TSP concentrations in Illinois. Figure 13. Trends in 24-hr maximum PM10 concentrations in Illinois. Figure 14. Trends in annual mean PM10 concentrations in Illinois. Figure 15. Trends in 24-hr maximum NO3- concentrations in Illinois. Figure 16. Trends in annual mean NO3- concentrations in Illinois. Figure 17. Trends in 24-hr maximum SO42- concentrations in Illinois. Figure 18. Trends in annual mean SO42- concentrations in Illinois. Figure 19. Trends in 24-hr maximum As concentrations in Illinois. Figure 20. Trends in annual mean As concentrations in Illinois. Figure 21. Trends in 24-hr maximum Cd concentrations in Illinois. Figure 22. Trends in annual mean Cd concentrations in Illinois. Figure 23. Trends in 24-hr maximum Cr concentrations in Illinois. Figure 24. Trends in annual mean Cr concentrations in Illinois. Figure 25. Trends in 24-hr maximum Fe concentrations in Illinois. Figure 26. Trends in annual mean Fe concentrations in Illinois. Figure 27. Trends in 24-hr maximum Mn concentrations in Illinois. Figure 28. Trends in annual mean Mn concentrations in Illinois. Figure 29. Trends in 24-hr maximum Ni concentrations in Illinois. Figure 30. Trends in annual mean Ni concentrations in Illinois. Figure 31. Trends in airborne metal enrichment, relative to soil, in Illinois. Figure 32. Spatial distributions of 1-hr maximum CO (ppm) in the Chicago area for 1980, 1985, and 1990. Figure 33. Spatial distributions of 8-hr maximum CO (ppm) in the Chicago area for 1980, 1985, and 1990. The shaded area denotes violation of an air quality standard. Figure 34. Spatial distributions of annual mean Pb (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 35. Spatial distributions of 24-hr maximum NO2 (ppm) in the Chicago area for 1980, 1985, and 1990. Figure 36. Spatial distributions of annual mean NO2 (ppm) in the Chicago area for 1980, 1985, and 1990. Shaded areas denote violations of an air quality standarD. Figure 37. Spatial distributions of 1-hr maximum O3 (ppm) in the Chicago area for 1980, 1985, and 1990. Shaded areas denote violations of an air quality standarD. Figure 38. Spatial distributions of 3-hr maximum SO2 (ppm) in the Chicago area for 1980, 1985, and 1990. Figure 39. Spatial distributions of 24-hr maximum SO2 (ppm) in the Chicago area for 1980, 1985, and 1990. The shaded area denotes violation of an air quality standarD. Figure 40. Spatial distributions of annual mean SO2 (ppm) in the Chicago area for 1980, 1985, and 1990. Figure 41. Spatial distributions of 24-hr maximum TSP (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 42. Spatial distributions of annual geometric mean TSP (mg/M3) in the Chicago area for 1980, 1985, and 1990. Shaded areas denote violations of an air quality standarD. Figure 43. Spatial distributions of 24-hr maximum NO3- (mg/M3) in the Chicago area for 1985 and 1990. Figure 44. Spatial distributions of annual mean NO3- (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 45. Spatial distributions of 24-hr maximum SO42- (mg/M3) in the Chicago area for 1985 and 1990. Figure 46. Spatial distributions of annual mean SO42- (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 47. Spatial distributions of 24-hr maximum Fe (mg/M3) in the Chicago area for 1985 and 1990. Figure 48. Spatial distributions of annual mean Fe (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 49. Spatial distributions of 24-hr maximum Mn (mg/M3) in the Chicago area for 1985 and 1990. Figure 50. Spatial distributions of annual mean Mn (mg/M3) in the Chicago area for 1980, 1985, and 1990. Figure 51. Spatial distributions of annual mean As, Cd, Cr, and Ni (mg/M3) in the Chicago area for 1990.

Return to Table of Contents