![[logo] US EPA](https://webarchive.library.unt.edu/web/20081106021705im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Attenuation of Ultraviolet Solar Radiation by Cloudy Skies: Links to Urban Air Quality
EPA Grant Number: CR830890C003Subproject: this is subproject number R830890C003 , established and managed by the Center Director under grant CR830890
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: Center for Environmental Science
Center Director: Frederick, John
Title: Attenuation of Ultraviolet Solar Radiation by Cloudy Skies: Links to Urban Air Quality
Investigators: Winiecki, Shelby , Frederick, John
Institution: University of Chicago
EPA Project Officer: Winner, Darrell
Project Period: July 1, 2003 through June 30, 2004 (Extended to June 30, 2006)
RFA: Targeted Research Center (2004)
Research Category: Hazardous Waste/Remediation , Targeted Research
Description:
Objective:This project examined a dataset of ultraviolet solar spectral irradiance acquired by a Brewer Spectrophotometer located in Chicago, Illinois. The objective was to assess the role of local cloudiness in controlling ground-level irradiance at wavelengths less than 350 nm, including the region of strong absorption by ozone at wavelengths shorter than 315 nm. The possibility that the attenuation associated with cloudy skies varies with wavelength was of particular interest.
Summary/Accomplishments (Outputs/Outcomes):The spectrophotometer was one of approximately 20 such instruments that comprised the U.S. Environmental Protection Agency’s (EPA) monitoring network. The instrument scanned the wavelength region from 286 nm to 363 nm, where the quantity sensed was the downward solar irradiance on a horizontal surface including the direct and diffuse components, at a spectral resolution of approximately 0.6 nm. The research reported here considers data collected from May through September 2001, an interval that spans the period of the year when photochemical processes are most likely to create degraded urban air quality.
We selected measurements in four wavelength ranges for detailed analysis. These are the “305 nm band”, defined as irradiance integrated over the range 303-308 nm, the “310 nm band” extending from 308-313 nm, the “315 nm band” from 313-318 nm, and the “345 nm band”, encompassing 340-350 nm. We denote these measured irradiances by EM(λ,θ) in watts m-2 where λ = 305, 310, 315, and 345 nm and θ is the solar zenith angle (SZA). Absorption by ozone is a major influence on ground-level irradiance in the shortest wavelength band and decreases to become negligible in the longest spectral interval.
To characterize the degree of cloudiness, we define the “transmission ratio” at 345 nm as:
T(345) = EM(345,θ)/EC(345,θ) (1)
where EC(345,θ) is the irradiance that would have existed under a clear sky at the time and SZA of the 345 nm measurement in the numerator. Obviously, T(345) = 1.0 corresponds to a clear sky, while a sky covered by thick clouds could have a value of 0.1 or less. To evaluate equation (1) we estimate EC(345,θ) by performing a smooth fit of the form EC(345,θ) = E0 exp[-τ/(cos θ)] to a set of points chosen to represent clear skies and determining the coefficients E0 and τ. The points selected were the ten largest irradiances measured in bins of width 0.05 in cos θ. The important outcome is to obtain values of T(345) that respond only to changes in cloudiness and not to changes in SZA.
A statistical regression model designed to detect wavelength dependence in the attenuation provided by cloudy skies must also account for absorption by ozone. An expression that accomplishes this is:
EM(λ,θ)/EM(345,θ) = T(345)ε(λ) exp[α(λ) - β(λ)Ω/(cos θ)] (2)
or:
ln[EM(λ,θ)/EM(345,θ)] = α(λ) - β(λ)Ω/(cos θ) + ε(λ) ln[T(345)] (3)
where Ω is total column ozone in DU and α(λ), β(λ) and ε(λ) are coefficients to be estimated. The mathematical form of equation (3) relates irradiance to column ozone, SZA and cloudiness, while still being expressible as a linear regression model. The dependence on column ozone and SZA is compatible with theory. If ε(λ) = 0, the model forces the fractional attenuation due to sources other than slant path column ozone to be independent of wavelength. A value of ε(λ) different from zero signifies wavelength dependent attenuation, where ε(λ) > 0 implies a greater attenuation at λ than at 345 nm.
Fits of equation (3) to the irradiance database for 305, 310, and 315 nm produced values of α(λ), β(λ), ε(λ), and their standard deviations. All of the regression coefficients were significantly different from zero at a confidence level higher than 99 percent. Second, the dependence of the irradiance ratio on slant path column ozone, expressed by β(λ), was consistent both in magnitude and wavelength dependence with expectations based on radiative transfer calculations. Finally, the derived values of ε were positive and indicate that, in the database as a whole, clouds provided a greater attenuation at 310 nm and 305 nm than at 345 nm. Furthermore, inclusion of the term that allows for a spectral dependence explains an increasing fraction of the variance as wavelength shrinks.
Since this research focuses on processes in the troposphere, it is desirable to develop a measure of ultraviolet irradiance that, for practical purposes, is independent of SZA and stratospheric ozone amounts. The transmission ratio, defined previously for 345 nm, can be generalized to all wavelengths studied:
T(λ) = EM(λ,θ)/EC(λ,θ,Ω) (4)
where EC(λ,θ,Ω) is the irradiance that would have existed under clear skies at the time of the measurement, and Ω is the column ozone amount. We used a radiative transfer model with column ozone measurements from TOMS to compute a value of EC(λ,θ,Ω) that pairs with each measured irradiance in the spectrophotometer’s dataset. The transmission ratios, T(345), T(315), T(310) and T(305), with small adjustments to remove identified biases, form the basis of the analysis, where emphasis is placed on the longest and the shortest wavelength. Attenuation by clouds and aerosols alone determine the values of T(345). The same processes influence T(305), but absorption by lower tropospheric ozone, that may differ from that implicit in the TOMS column value used to compute EC(λ,θ,Ω), is also a factor.
Figure 1 depicts the observed relationship between transmission ratios for the 305 nm and 345 nm bands derived from each spectrophotometer scan, where logarithmic scales highlight behavior at small values, which correspond to dense clouds. The plus signs in Figure 1 define the case where no wavelength dependence exists, ln[T(305)] = ln[T(345)]. The open circles show the inequality T(305) < T(345) to be typical and that T(305) shrinks by a disproportionately large amount as T(345) decreases. The important information in Figure 1 arises from the nonlinear relationship between data at different wavelengths. This nonlinearity is unaffected by any normalization constants applied to the transmission ratios to remove biases from equation (4).
Figure 1. Transmission Ratios in the 305 nm Band Plotted as a Function of Corresponding Values for the 345 nm Band for all Points in the Dataset (Open Circles). The axes are expressed as natural logarithms to highlight the nonlinear behavior at small values, which correspond to thick cloud cover. Plus signs define the case ln[T(305)] = ln[T(345)].
The apparent excess attenuation at 305 nm (as well as at 310 nm, not shown here) could arise from a problem in the analysis, or it may be a consequence of atmospheric processes. We examined the first of these possibilities, emphasizing the 305 nm band. Satellite-based column ozone values enter Figure 1 through the computed clear-sky irradiance in the denominator of the transmission ratio, equation (4). The backscattered radiances measured by TOMS are sensitive to both the ozone amount and the lower boundary albedo in the instrument’s field of view, particularly under conditions of thick cloud cover that produce small values of ln[T(345)]. If an error existed in the reported column ozone, and the magnitude of this error is correlated with the albedo, then the nonlinear behavior in Figure 1 might result. Specifically, if the column ozone value deduced from TOMS over dense cloud cover was smaller than the true value, then the computed transmission ratio at 305 nm would be too small as well, while that at 345 nm would be unaffected. We have no information to suggest that such an error indeed exists in the satellite-based results, but it is important to examine this option before seeking a physical explanation for the behavior in Figure 1. We simulated transmission ratios for a range of cloud albedos and hypothetical errors in column ozone amounts. Based on analysis of these results, we concluded that reasonable errors in the column ozone values used to generate clear-sky irradiances are unable to account for the behavior in Figure 1.
Absorption by ozone located in the lower troposphere, to which the TOMS measurements are insensitive, remained a potential explanation for the characteristics in Figure 1. This was a promising option because the computed clear-sky irradiances used in equation (4) do not include enhanced ozone amounts typical of an urban boundary layer. As an initial test for a relationship, we fit a regression model of the form ln[T(305)] = a0 + a1ln[T(345)] + a2χ to the database, where χ is the surface ozone mixing ratio measured on the rooftop adjacent to the spectrophotometer, and a0, a1 and a2 are determined from the fit. The resulting value of a2 was negative with very high statistical significance.
Motivated by the above correlation, we performed a series of radiative transfer calculations that varied both cloudiness and the ozone mixing ratio in a 2-km thick layer beneath the cloud, adjacent to the ground. The objective was to determine if clouds located above a boundary layer containing enhanced ozone amounts could produce the behavior in Figure 1. The results showed that even a boundary layer mixing ratio of 90 ppbv, the largest recorded during the period of observation, could not lead to differences between ln[T(345)] and ln[T(305)] of the magnitude shown by the spectrophotometer data. Despite this negative result, the correlation between ln[T(305)] and surface ozone amounts suggests a link between tropospheric air quality and the characteristics of Figure 1. An acceptable physical mechanism must produce the inequality ln[T(305)] < ln[T(345)] and nonlinear behavior in the range -3 < ln[T(345)] < -2, which corresponds to thick cloud cover.
The last alternative was to consider wavelength-dependent attenuation when ozone is placed in the cloud itself. When we simulated the effect of reasonable ozone mixing ratios in the interstitial air of a cloud, the results displayed a nonlinear relationship that agreed extremely well with Figure 1. When ln[T(345)] = -3, the simulation yields ln[T(305)] = -3.48 and -4.27 for ozone mixing ratios of 30 and 90 ppbv, respectively. Furthermore, the computed values appear nearly linear over the range -2 < ln[Tc(345)] < 0, with nonlinearity developing when ln[Tc(345)] < -2. Based on these results, it is reasonable to attribute the behavior deduced empirically to absorption by ozone contained within the clouds observed by the spectrophotometer. While the above explanation produces results consistent with the observations, it does not constitute a proof that ozone alone is the causal agent. Specifically, we cannot disregard the possibility of wavelength dependent attenuation associated with particulate matter or species dissolved in cloud water.
Conclusions:Measurements from the Chicago-based Brewer Spectrophotometer combined with radiative transfer calculations produce transmission ratios in the 305 nm wavelength band that are generally smaller than those for the 345 nm band under the most dense clouds. Furthermore, results for the 310 nm and 315 nm bands (not shown here) indicated a smooth wavelength dependence in this excess absorption. The nonlinear behavior in Figure 1 implies that the observed differences could not result from fixed, but wavelength-dependent, offsets in the computed transmission ratios.
A simple model of absorption and scattering within a cloud produced results that were consistent with the observations. These reveal a likely coupling between the ultraviolet transmission properties of clouds and the ozone content of the troposphere. The combination of a gaseous absorber and a medium that is optically thick in scattering leads to a greater attenuation than when the same quantity of absorber exists in clear air. The observed excess attenuation near 305 nm, if it is the result of absorption by ozone, indicates enhanced production of O(1D), and therefore altered photochemical activity, in clouds.
Journal Articles on this Report: 1 Displayed | Download in RIS Format
| Other subproject views: | All 2 publications | 1 publications in selected types | All 1 journal articles |
| Other center views: | All 9 publications | 2 publications in selected types | All 1 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Winiecki S, Frederick JE. Ultraviolet radiation and clouds: couplings to tropospheric air quality. Journal of Geophysical Research 2005;110(D22):D22202, doi:10.1029/2005JD006199. |
CR830890C003 (Final) |
not available |
ultraviolet radiation, brewer spectrophotometer, clouds, ozone, air quality,
,
Air, Scientific Discipline, Health, RFA, Engineering, Chemistry, & Physics, Risk Assessments, Health Risk Assessment, Biochemistry, particulate matter, Ecology and Ecosystems, solar irradiance, cardiovascular disease, cardiovascular vulnerability, air quality, ozone, human health risk, urban air , ultaviolet illumination, human health effects, particulates, PM 2.5, air pollution, airway disease, vascular dysfunction, ultraviolet radiation, airborne particulate matter
Relevant Websites:
Progress and Final Reports:
Original Abstract
Main Center Abstract and Reports:
CR830890 Center for Environmental Science
Subprojects under this Center:
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
CR830890C001 The Urban Measurements Project—The Urban Atmosphere Observatory
CR830890C002 The Urban Data Analysis and Modeling Project
CR830890C003 Attenuation of Ultraviolet Solar Radiation by Cloudy Skies: Links to Urban Air Quality
CR830890C004 Measurements of Black Carbon in Chicago: Implications for Controls on Diesel Emissions
CR830890C005 Attenuation of Visible Sunlight by Limited Visibility and Cloudiness
CR830890C006 The Energy Balance of Urban Microclimates