Study

A study by Barnard (2001) shows that specific air pollutants can reduce the increased surface levels of UV radiation, and offers an explanation for why the expected surface UV increases, caused by decreases in stratospheric ozone, have not been observed, especially in urban regions. Air pollution data (NOX, O3, PM-10, SO2) collected from an EPA UV monitoring site at the University of California at Riverside (UCR), combined with data from a site operated by the California Air Resources Board in Rubidoux, CA, provided the basis of this study. The 1997 South Coast Ozone Study (SCOS) provided three key ingredients: (1) black carbon concentrations, (2) PM-10 concentrations, and (3) collocated radiometric measurements. The Total Ozone Mapping Spectrometer (TOMS) satellite data were used to provide the stratospheric ozone levels that were included in the statistical model. All of these input parameters were used to test this study's hypothesis: "the expected increase of surface UV radiation, caused by decreases in stratospheric ozone, can be masked by increases in anthropogenic emissions." The values for the pollutants were 7:00 a.m. - 5:00 p.m. averages of the instrument's values taken during the summer of 1997. A statistical linear regression model was employed using the stratospheric ozone, black carbon, PM-10, and surface ozone concentrations, and the sin (0) and cos (0). The angle (9 is defined by: 0 = 2n (Julian date/365). This model obtained a coefficient of determination of 0.94 with an uncertainty level (p-value) of less than 0.3% for all of the variables in the model except ground-based ozone. The final model, regressed against a data set from a remote, western North Carolina site, resulted in a coefficient of determination of 0.92. The model shows that black carbon can reduce the Diffey weighted UV (DUV) levels that reach the surface by as much as 35%, depending on the season and location.

After review of the various pollutants that absorb in the UV range, only the possibility of PM-10 and a subset of PM-10, the black carbon, and ground-based ozone, were found in high enough concentrations to warrant further investigation. These high concentrations, especially of black carbon, that were found in this urban environment of Riverside-Rubidoux, CA, had not been previously experienced by NC State University personnel who conducted research in the mountains of North Carolina (Im et al., 2001).

Graphs of DUV changes and corresponding stratospheric ozone changes were reviewed for the entire SCOS-97 study. Three days (September 7 - 9), were chosen to study behavioral changes with respect to TOMS ozone, DUV, and black carbon. The graph for this period is shown in Fig. 11.11. Along with the DUV and TOMS, data curves represent the black carbon values for each day of interest. For days 6 - 9, the TOMS ozone value dropped equivalent amounts between days 6 and 7, and between days 8 and 9. However, the DUV increased by over 300 joules m from day 6 to day 7, but remained constant from day 8 to day 9, even with a TOMS ozone decrease. In this particular instance, the averaged black carbon for days 6 and 7 is 1.29 ^g m , while the averaged black carbon for days 8 and 9 is 2.02 ^g m , a substantial increase. Figure 11.11 also represents the first set of data that are used to relate black carbon and DUV decreases.

Days of month

* Numbers oil the graph represent black carbon 7:00 a.m. - 5:00 p.m. average in micrograms per cubic meler.

Figure 11.11 Brewer DUV and TOMS Ozone, Sept. 1997

Days of month

* Numbers oil the graph represent black carbon 7:00 a.m. - 5:00 p.m. average in micrograms per cubic meler.

Figure 11.11 Brewer DUV and TOMS Ozone, Sept. 1997

A Statistical Analysis System (SAS) linear regression model was employed that enabled more of the SCOS-97 data to be included. Approximately 63 clear sky days, determined with Brewer DUV readings and the UCR visible data, were investigated.

Ground-based ozone, PM-10, black carbon, and TOMS ozone were used as the independent variables that would be modeled to make up the dependant DUV variable. The model included two other parameters, the sin (0) and cos (0), to account for DUV seasonal variations caused by the earth's rotation around the sun and the tilt of its axis, where 0 = 2n ((Julian day) x (365-1)). Maximum DUV levels at the UC Riverside site occurred during the summer months reaching close to 5,000 joules m-2 day-1 and minimum levels (781.6 joules m-2 day-1) occurred during the winter months.

The results of the SAS linear model are presented below. The final regression developed from this data set is:

DUV = 608.91 sin (0) -2,049cos (0) -16.93 xTOMS - 7.1

TOMS ozone values are in Dobson units. The units for PM-10 and black carbon are micrograms per cubic meter (^g m-3).

One must be careful at this point in interpreting the results. The ¿-statistics employed here assume a normal distribution, centered on zero with tails in the positive and negative direction. Physically, however, the coefficients for black carbon, PM-10, TOMS, and ground-based ozone must be negative, as all are suspected to be UV radiation absorbers. If positive coefficients had been obtained for any of these independent variables, it would have implied UV radiation had been created, which is not physically possible.

Table 11.2 indicates the sensitivities of each of the parameters used in the model. Table 11.3 was developed using the maxima and minima parameters from this study, and average values based on the author's previous experience in the pollution-monitoring field. The changes noted in Tables 11.2 and 11.3 correlate well with expected physical results. As the Julian day changes from summer (Julian day 180) to winter (Julian day 365), the surface DUV changes dramatically (4,122.3 joules m-2). Likewise, the changes brought about by the TOMS ozone are in line with current literature. A model check done with the radiation amplification factor (RAF) shows excellent agreement with accepted RAF values. This value is known (Madronich et al., 1998) to be between 1.2 and 1.3, and is defined by the percent change in UV divided by the percent change in TOMS ozone. If the change in UV, or in this case DUV, is divided by the mean of the DUV values, a percent change is obtained. Thus, using the appropriate values and the statistical model, the RAF =16.93 x 294.5/4,036.5 =1.235, confirming the relationship between the DUV and TOMS ozone. The surprising factors are the coefficients for the black carbon and the PM-10. Using the values in Table 2, the changes of black carbon and PM-10 encountered in this research account for 46.5% and 58.5%, respectively, of the DUV changes caused by TOMS ozone.

One pollutant, which was not accounted for at all in Eq. (11.1), was the ground-based ozone. Long thought to be one of the significant absorbers of UV-B radiation, it did not statistically meet the 0.05 significance level in this model. Several trials of the SAS Autoreg procedure, that included non-linear fits of the data and SIN and COS squared terms, did not produce models with coefficients of determination as high as for Eq. (11.1), nor did they produce coefficients that were physically meaningful. When this is carefully analyzed, it makes sense. UV radiation is required to generate the surface measured ozone, but in return, the more ozone, the less UV radiation penetrates to the surface. Also, the more black carbon, the less the UV, and thus the reduction of surface measured ozone.

One factor in the model, which dominates the equation, is the intercept (8,484 joules m-2 day-1). This number, much larger than the other terms of the equation, needs some physical explanation. Mathematically, it is the value of the DUV radiation, dependant variable (Y), when all other independent variables (Xs) are zero. Physically it represents a daily averaged DUV value found at the top of the atmosphere. If all the independent variables are removed from the equation, the radiation has no atmospheric constituents to attenuate it and no day-to-day variations caused by the earth's tilting axis and orbit around the sun. If the DUV changes created by these independent variables (Xs) are added to the maximum DUV level from Table 11.2, they sum to 7,994.3 joules m-2 day-1. This would approximate a lower boundary level for the daily averaged DUV at the top of the atmosphere. If the same is done for Table 11.3, they sum to 11,481.9 joules m-2 day-1. These two numbers represent the range of variations in the DUV values at the top of the atmosphere caused by the maxima and minima values of the factors which alter the radiation as it passes from the top of the atmosphere to the surface. The intercept (8,484 joules m-2 day-1) for this model lies within this range. Thus, the intercept could be interpreted as a constant similar to that of the I0 in the Langley expressions for optical depth.

While this regression was developed utilizing data from a large West Coast experiment in a highly polluted area, it was tested and confirmed in a more pristine, albeit still polluted, East Coast remote site. The Mt. Gibbes site in Mt. Mitchell State Park, NC has been operating for some time with various types of aerosol and radiation equipment. The data from the spring, summer, and fall of 1999 included UV solar radiation and black carbon. TOMS ozone values were obtained from NASA. Since no collocated data were available for the PM-10 concentrations, a yearly averaged value from a neighboring county was used. These were then input into Eq. (11.1) and approximately 30 values of DUV were generated. The values for black carbon at this site ranged from a low of 0.032 ^g m-3 to a high of 1.545 ^g m , considerably less than those found at the Riverside site (Table 11.4). UVB-1 data set values, taken at one-minute intervals, were added to obtain a daily (7:00 a.m. - 5:00 p.m.) total UV exposure. This value was then scaled by the CIE (Diffey erythemal curve) factor described in the operations manual for the UVB-1 instrument. The model results were then plotted against these 7:00 a.m. - 5:00 p.m. values.

Table 11.2 Sensitivity tests of DUV in Eq. (11.7) for expected ambient maxima and minima values; the tests consider the parameters of (a) Julian date, (b) TOMS ozone, (c) PM-10, and (d) Black carbon, as the real variables each time, while the others (1999 yearly means) were treated as constants

(a)

DUV

Julian Day

TOMS Ozone

PM-10

Black Carbon

(J m 2 d ^

(DU)

(^g m 3)

(ng m 3)

777

1

3,380

90

4,889

180

294

65

1,546

2,340

270

767

365

Change

4,122

(b)

DUV

TOMS Ozone

Julian Day

PM-10

Black Carbon

(J m 2 d 1)

(DU)

(^g m 3)

(ng m 3)

3,031

402

183

65

1,546

5,723

243

Change

-2,692

159

(c)

DUV

PM-10

Julian Day

TOMS

Black Carbon

(J m 2d 1)

m 3)

Ozone (DU)

(ng m 3)

4,164

163

183

294

1,546

5,271

7

Change

-1,107

156

(d)

DUV

Black Carbon

Julian Day

TOMS

PM-10

(J m 2d 1)

(ng m 3)

Ozone (DU)

(^g m 3)

4,050

7,000

183

294

65

5,087

10

Change

-1,037

6,990

The most striking findings are the results from this testing. As shown in Fig. 11.12, where the modeled data is plotted against the measured UV data from the UVB-1 instrument, there is an excellent correlation, R2 = 0.92. In spite of utilizing one type of instrument to develop the model (Brewer spectrophotometer), and comparing the model's output to that of another type of instrument (Yankee UVB-1), integrated DUV radiation values between the two sites, and the instruments, correlated extremely well. The model (Eq. (11.1)) was applicable for black carbon values ranging from 0.03 ^g m to above 4.0 ^g m , a range of over two decades. It maintained its linearity over this entire range. The Yankee UVB-1 instrument does not have a spectrally resolved output. It measures an entire UV spectrum that includes both the UV-B region and the UV-A region.

Table 11.3 Sensitivity tests for the independent variables of Eq. (11.7) for expected ambient minima and maxima: (a) Julian date, (b) TOMS ozone, (c) PM-10, and (d) Black carbon

Table 11.3 Sensitivity tests for the independent variables of Eq. (11.7) for expected ambient minima and maxima: (a) Julian date, (b) TOMS ozone, (c) PM-10, and (d) Black carbon

DUV

Julian Date

TOMS PM-10 Ozone

Black Carbon

781.6

0

300 DU 60 ^g m

3 1.0 ^g m 3

3,395.3

90

4,903.9

180

2,355.1

270

781.6

365

Change

4,122.3 joules m 2

day 1

(b)

DUV

TOMS Ozone

PM-10

Black Carbon

2,364.4

450

60 m 3

1.0 ^g m 3

6,173.6

225

Change

-3,809.2 joules m

2 day 1

225 DU

(c)

DUV

PM-10

TOMS Ozone

Black Carbon

3554.9

250

300 DU

1.0 ^g m 3

5294.4

5

Change

-1739.5 joules m

2 day 1

245 ^g m 3

(d)

DUV

Black Carbon

TOMS Ozone

PM-10

5,050.8

7.00

300 DU

60 m 3

4,013.5

0.01

Change

-1,037.3 joules m

2 day 1

6.99 ^g m 3

Table 11.4 Maxima, minima, mean, and standard deviations for the values of the variables used to develop and test the model

Riverside, CA Mt. Gibbes, NC

Table 11.4 Maxima, minima, mean, and standard deviations for the values of the variables used to develop and test the model

Riverside, CA Mt. Gibbes, NC

Max.

Min.

Mean

Std. Dev.

Max.

Min.

Mean

Std. Dev.

DUV* (J m 2d 1)

6,035

2,038

3,698

990

9,677

3,257

6,257

1,764

TOMS (DU)

330

259

285

15

348

270

298

20

PM-10 (^g m3)

132

15

47

20

60**

7

28

12

Black Carbon (ng m

3) 4,757

181

1,546

761

1,545

32

395

366

Ozone*** (ppb)

105

25

59

18

na

na

na

na

* UV values obtained from a Brewer spectrophotometer at Riverside, CA and integrated UVB-1 values from Mt. Gibbes, NC.

** PM-10 values from EPA/AIRS database for 1999, yearly average value of 28 used to test model. *** Ground-based ozone values were used to determine the SAS model, but were not used with the Mt. Gibbes data. Their significance was above the .05 level.

Pollutant concentrations are averages of the hourly averages from 7:00 a.m. -5:00 p.m.

* UV values obtained from a Brewer spectrophotometer at Riverside, CA and integrated UVB-1 values from Mt. Gibbes, NC.

** PM-10 values from EPA/AIRS database for 1999, yearly average value of 28 used to test model. *** Ground-based ozone values were used to determine the SAS model, but were not used with the Mt. Gibbes data. Their significance was above the .05 level.

Pollutant concentrations are averages of the hourly averages from 7:00 a.m. -5:00 p.m.

_ 12000

t ioooo

I 6000

g 4000

1500 2000 2500 3000 3500 4000 4500 Figure 11.12 UVB-1 Mt. Gibbes daily integrated values

Other studies (Liousse, et al., 1996) have concluded that the black carbon could act to cool the earth, yet UV radiation, being in the shorter end of the spectra, has the highest energy photons. Since the black carbon is absorbing these high-energy photons, and probably re-emitting in the infrared, it would be a logical conclusion that the atmospheric layer in which black carbon resides would be slightly heated. While the UV heating factor would be small, the black carbon absorbs strongly at all wavelengths and can act as a global cooling factor, reducing the available radiation to the earth's surface (Satheesh and Ramanathan, 2000; Yu et al., 2001). The heating of the atmosphere could also impact the photochemical reactions and cloud formations (Ackerman et al., 2000) that occur. There would be a trade-off in this area. The black carbon is reducing the UV that is available to initiate photochemical reactions, but black carbon's presence will also heat the surrounding atmosphere, which tends to speed up these same reactions.

In this particular urban environment, the author found that the black carbon concentrations seem to have little effect on the aerosol optical depth, but did show a relationship with the single scatter albedo. Two papers (Lioussie et al., 1996; Dubovik et al., 1998) that discuss black carbon's optical properties focus on its relationship with the single scatter albedo and not on the optical depth.

In this study statistics and monitoring data led to the formulation of a model that determined an average daily UV dosage at the surface. PM-10 and black carbon were observed to have a significant impact on the UV, but this model still needs to be tested in other regions to refine it with more definitive particulate measurements (chemical composition and size). The research from this study should lead to further investigations that may help develop more quantitative means that would be useful in describing the effects of aerosols on the radiative transfer of UV to the surface. Further work can be done to determine if the long-term increases of automobile and truck emissions are capable of producing the possible effects shown here and to what extent their increases correlate to the decreases of UV radiation. Black carbon's direct and indirect effects on the photochemical

*

y=2.

46a- 807.: /?-=0.92

>9

processes in the atmosphere also need to be fully investigated with studies that encompass varied regions.

Was this article helpful?

0 0

Post a comment