![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081107184900im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Surface Levels of Ultraviolet-B Radiation Under Variable Conditions of Tropospheric Air Quality And Cloudiness
EPA Grant Number: R825248Title: Surface Levels of Ultraviolet-B Radiation Under Variable Conditions of Tropospheric Air Quality And Cloudiness
Investigators: Saxena, Vinod K. , Frederick, John
Institution: North Carolina State University , University of Chicago
EPA Project Officer: Shapiro, Paul
Project Period: October 1, 1996 through September 30, 1999
Project Amount: $374,702
RFA: Exploratory Research - Air Chemistry & Physics (1996)
Research Category: Engineering and Environmental Chemistry
Description:
Objective:Ultraviolet radiation (UV) is a driving factor in atmospheric photochemical processes, and an increase in surface UV irradiance could lead to damaged vegetation and increases in incidences of skin cancer and cataracts. The overall objective of this research project was to provide a pilot database for interpreting the measurements of surface UV-B radiation under variable conditions of tropospheric air quality and cloud cover. Analysis of the database was to help answer the following question: How successfully can we detect a downward trend in stratospheric ozone by monitoring the surface UV-B irradiance from a network of ground-based stations in the United States? Further analysis was to help refine the forecast of the UV-B Index (UVI) and refine inputs to ultraviolet radiative transfer models.
Two sites, 1 kilometer vertically and 10 kilometers horizontally apart, situated in the Blue Ridge Mountains of western North Carolina, were outfitted with instrumentation beginning in June 1997. The instrumentation was installed as it became available to us from North Atlantic Assembly-Serb and the U.S. Environmental Protection Agency's Atmospheric Research and Exposure Assessment Laboratory (EPA-AREAL). The experimental site and instrumentation are described in detail by Schafer, et al. (1996), and Wenny, et al. (1998). The instruments deployed relevant to this project are the Yankee Multi-Filter Rotating Shadowband Radiometer (MFRSR), Yankee Ultraviolet Multi-Filter Radiometer (UVMFR), Yankee UVB-1 Radiometer (UVB1), Brewer Spectrophotometer (Brewer), TSI Differential Mobility Particle Sizer (DMPS), TSI Integrating Nephelometer (TSIN), Radiance Research Integrating Nephelometer (RRIN), and a Magee Scientific Aethalometer. The MFRSR provides simultaneous measurements of global, diffuse, and direct components of solar irradiance at six wavelengths (415, 500, 615, 673, 870, and 940 nm). The UVMFR provides simultaneous measurements of global, diffuse, and direct components of solar irradiance at seven wavelengths (300, 305.5, 311.5, 317.5, 325, 332.5, and 368 nm). The UVB1 provides measurements of broadband UV-B global irradiance (280-320 nm). The Brewer provides measurements of both total ozone column (TOC) and spectral UV global irradiance (286.5-363 nm in 0.5 nm increments). The TSIN measures total scattering and backscattering coefficients at three wavelengths (450, 550, and 700 nm). The RRIN measures total scatting coefficient at 530 nm. Backward air parcel trajectories were computed using the Hybrid Single Particle Lagrangian Integrated Trajectory (HY-SPLIT) model. The criteria for designating air masses as P, C, and M are based on the emission inventories of the EPA An Aethalometer manufactured by Magee Scientific was used to make black carbon (BC) measurements. The TSI DMPS measures particle size distributions (0.016-0.6 m).
Aerosol optical depth (AOD) was derived from the MFRSR direct irradiance measurements using the Bouguer-Langley method (Lenoble, 1993). UV-B transmission over the layer defined by the two sites was determined by the ratio of the UVB1 measurements at the two sites. The ground albedo and the imaginary part of the refractive index were calculated using a unique procedure involving a Mie code and a radiative transfer code in conjunction with the retrieved columnar aerosol size distribution, AOD, and diffuse-direct ratio known as the Search Graph Method.
During November 1997, an intensive field campaign was conducted in conjunction with researchers from the University of Miami. The focus of this campaign was the collection of aerosol samples using optical sizing instruments (TSI Scanning Mobility Particle Sizer, TSI Ultra-Fine Particle Counter, TSI Condensation Particle Counter, and TSI Aerodynamic Particle Sizer).
At present, data acquisition continues at both the mountain and valley sites. The findings presented here contain an analysis of data from 1995 to 1999.
Summary/Accomplishments (Outputs/Outcomes):The clear-sky broadband UV-B transmission over the layer defined by the two sites was determined for a number of days with contrasting air mass histories. The average UV-B transmission was 77.8 ± 4.6 percent, 81.9 ± 2.9 percent, and 83.5 ± 2.8 percent for polluted, marine, and continental air masses, respectively. Our observed range of broadband UV-B attenuation of 14 to 31 percent for the 1-kilometer layer is consistent with other reported values of transmission through 1 kilometer in the lower troposphere. The observed differences for layer transmission for the three air mass types are attributed to the contrasting aerosol properties of the three air mass types.
The Brewer spectral UV measurements from the valley site were analyzed to determine the wavelength-dependent effects of clouds and aerosols on solar transmission. Normalizing the irradiance measurements to a clear-sky measurement, it was found that the normalized irradiance increased with wavelength on certain days and decreased with wavelength on other days. This is consistent with the modeled results of Erlick and Frederick (1998a, 1998b), indicating that aerosols in cloudy atmospheres can cause normalized transmission to increase with wavelength in the ultraviolet and visible, while cleaner clouds cause normalized transmission to decrease with wavelength from 320 nm through the visible.
The measurement of the physico-chemical properties of aerosols in the atmospheric
layer between the two sites was made to estimate the optical properties of the
intervening layer. During the months from July to December 1995, the total representative
optical depths at 500 nm at the valley site were 0.29 ± 0.19, 0.10 ±
0.04, and 0.68 ± 0.33 for M, C, and P air masses, respectively. Total
optical depths at 500 nm on Mt. Mitchell were 0.10 ± 0.09, 0.032 ±
0.01, and 0.19 ± 0.12 for M, C, and P air mass types, respectively. The
ratios of mean 1-kilometer layer optical depth between the mountain and the
valley to total mean optical depth at 500 nm from the valley site were 71 percent,
68 percent, and 73 percent for M, C, and P air masses, respectively. This indicates
that the major part of atmospheric aerosols is located within the lowest 1-kilometer
boundary layer of the troposphere. The average diffuse-direct ratios at the
valley site for P, M, and C air masses at 500 nm were 0.98 ± 0.33, 0.37
± 0.29, and 0.14 ± 0.07, respectively. There was a significant
linear correlation between the diffuse-to-direct ratio and the total AOD at
both the mountain and the valley site in conformity with theoretical predictions.
A search-graph method was developed (Yu, et al., 2001) and used to retrieve
the columnar lognormal size distribution (N, rg, and 
g)
on the basis of the optical depth at three wavelengths (415, 500, and 673 nm).
The ground albedo and imaginary part of the refractive index were calculated
using a unique procedure involving a Mie code and a radiative transfer code
in conjunction with the retrieved aerosol size distribution, AOD, and diffuse-to-direct
ratio. It was found that rg and g
were in the range of 0.022 to 0.18 µm and 1.37 to 2.95, respectively.
The asymmetry parameter and single scattering albedo were in the range of 0.607
to 0.729 and 0.849 to 0.989, respectively. The average ground albedo and imaginary
part of the refractive index were 0.17 and 0.019, respectively. Application
of these radiative transfer model techniques predicts a spectral dependence
of UV transmission through a cloudy atmosphere with an added dependence on the
amount of aerosols present. The Brewer spectrophotometer measurements of spectral
UV-B irradiance display evidence of the spectral dependence of UV transmission
as predicted by the modeling.
In the visible wavelengths, the AOD and diffuse-to-direct ratio exhibit a dependence on air mass type. In the UV wavelengths, the aerosol physico-chemical properties derived from the data also reveal dependencies on air mass type.
The role of BC aerosols on cloud microphysical/optical properties was investigated. Fifteen-minute BC mass concentrations have been obtained for clear air days in June-October 1996, and March-October 1997. In general, average BC mass concentrations from the P sector were higher than the C and M sectors. These results compare well with those obtained by Chylek, et al. (1996) in southern Nova Scotia and those found at Mace Head, Ireland by Jennings, et al. (1993). Average values of transAtlantic BC transport to Mace Head ranged from 7 to 21 ng m-3. The average BC data collected at Mt. Mitchell for M air was 62 ng m-3. The higher value can be expected for Mt. Mitchell because the marine air may have been modified by traversing over land before reaching the site.
The occurrence of a long-lasting forest fire in Mexico during May 1998, provided evidence that the research sites are indeed influenced by the long-range transport of aerosols. A smoke plume originating in Mexico was observed by the Total Ozone Mapping Spectrometer (TOMS) instrument, and subsequent images showed the plume passing over the Southeastern United States. A substantial increase in the BC mass concentration measurement at the mountain site was observed, coinciding with the passage of the plume (an increase on the order of seven times the pre-plume BC concentrations). Light scattering measurements at the valley site also showed a sharp increase coinciding with passage of the plume.
Researchers at the University of Chicago examined the behavior of ground-level solar UV irradiance measured from several locations during the 1990s. The data sets come from four Brewer spectrophotometers located in Japan (Naha, Kagoshima, Tateno, and Sapporo) and three Brewer spectrophotometers in Canada (Saturna, Winnipeg, and Montreal). The goal is to define the variability that occurs in UV irradiance over time scales ranging from days to several years. Since the duration of the data sets is 5 years or less, conclusions concerning systematic trends should not be made; however, any trends associated with a decline in or recovery of stratospheric ozone amounts is meshed with (and obscured by) the behavior identified in this work. Spectral information provided by the Brewer instruments allows the observed variability to be interpreted in terms of cause and effect. The 24-hour integrated irradiance at a wavelength of 322.5 nm and the erythemally-weighted irradiance both show day-to-day variations at ranges typically ± 30-35 percent of their monthly average values. This similarity in interday variability, specifically the absence of a wavelength dependence, eliminates fluctuations in ozone as the primary cause. A similar study of monthly integrated UV irradiances at selected wavelengths from 300-322.5 nm demonstrates that cloudiness is by far the major driver of interannual variability. This is true, in particular, of the biologically weighted irradiance for erythema. Changes in ozone from 1 year to the next are a secondary factor that account for approximately one-fifth of the interannual variance in erythemal irradiance for the sites and time period studied.
A conventional delta-Eddington radiative transfer code was modified to simplify the modeling of the highly anisotropic scattering properties of cloud droplets and aerosol particles in the atmosphere. The modification involves independent computation of the singly scattered and multiply scattered radiation, allowing the singly scattered radiation to be computed analytically. The standard Eddington approximation is applied to only the multiply scattered radiation component. This modified delta-Eddington model adequately handled strong absorption in the stratosphere and the anisotropic scattering within optically thin cloud layers, with results better than those from a simple two-stream or conventional delta-Eddington model.
The wavelength dependence of aerosols in cloudy atmospheres was investigated using the modified delta-Eddington model (Erlick, et al., 1998). When clouds are superimposed on an aerosol profile with the cloud drops and aerosol particles externally mixed, the shape of the normalized transmission spectrum is dominated by the effect of the cloud drops, unless the optical depth of the aerosols begins to approach the optical depth of the cloud, such as when an optically thin stratus cloud is superimposed on an urban aerosol profile. If cloud drops and aerosol particles are internally mixed through coagulation, the shape of the normalized transmission spectrum is again dominated by the cloud drops, unless there is an unrealistically high volume fraction of strongly absorbing inclusions inside the droplets. The number density of aerosol particles outside the cloud has a greater influence on the magnitude and shape of the transmission spectrum than variations in interstitial AOD in the volume fraction of absorbing inclusions in the cloud drops.
The Brewer spectral UV measurements from the valley research site were analyzed to determine the wavelength-dependent effects of clouds and aerosols on solar transmission. Normalizing the irradiance measurements to a clear-sky measurement, it was found that the normalized irradiance increased with wavelength on certain days and decreased with wavelength on other days. This is consistent with the modeled results discussed above, indicating that aerosols in cloudy atmospheres can cause normalized transmission to increase with wavelength in the ultraviolet and visible, while cleaner clouds cause normalized transmission to decrease with wavelength from 320 nm through 700 nm. The implications are that the attenuation of clouds in the UV cannot always be accurately estimated from visible sunlight data when aerosols are present. In continental or urban regions prone to high aerosol loadings, the attenuation of clouds with optical depths on the order of 10 or less may either increase or decrease with wavelength depending on atmospheric conditions. Measurements of ground-level irradiance at both UV and visible wavelengths may give an indication of the level of aerosol loading in the atmosphere under clear skies or in the presence of optically thin clouds. An intercomparison of total ozone measurements was conducted using two independent pairs of Brewer and Microtops instruments. The measurements demonstrated good agreement, with a mean percentage difference of ± 3 percent for both simultaneous individual measurements and daily averaged ozone column values. A comparison of total ozone (daily average) measurements as recorded by Brewer, Microtops II, and TOMS demonstrated a consistent trend in the magnitude of ozone column amounts. The consistency and quality of the agreement indicated that the Microtops may be a suitable instrument to deploy in the field for total column ozone measurements, as long as it is subject to periodic calibrations against an accepted standard.
Project research also has focused on developing methods and deriving the TOC
and AOD using the measurements of an UVMFR. The impact of these derived quantities
on surface erythemal UV levels and the National Weather Service UV Index (UVI)
was assessed. Surface measurements of total and diffuse UV irradiance at the
seven narrowband wavelength channels of the UVMFR were used to determine total
column ozone and AOD for two 6-month periods in 1997 and 1999. The derived TOC
displayed a seasonal pattern of higher column amounts during the summer and
lower amounts during the fall/winter as expected for the northern mid-latitudes.
A comparison with the ozone column derived from the TOMS showed a mean ratio,
UVMFR(O3)/TOMS(O3), of 0.98 (standard deviation = 0.02) for 1997 and 0.98 (standard
deviation = 0.01) for 1999 (TOMS data obtained from http://jwocky.gsfc.nasa.gov 
).
The repeatability in the ozone retrieval for clear periods demonstrates that
the UVMFR is an adequate field research tool for obtaining accurate ground-based
TOC measurements. The automated UVMFR takes measurements continuously and does
not require frequent operator attention, thus providing a greater ozone retrieval
rate than other ground-based instruments, such as the Dobson or Brewer Spectrophotometer.
It must be stressed, however, that regular external lamp calibrations are necessary
for the UVMFR to detect and correct any drift in the cosine response and spectral
response functions of the filters.
The AOD was derived from the UVMFR for all seven wavelength channels during significant clear sky times throughout the 6-month measurement period in 1999. The AOD at UV wavelengths is difficult to retrieve with great accuracy due to the complicating factors of strong ozone absorption and Rayleigh scattering in the UV region, as well as the relatively large inherent uncertainty in the UV irradiance measurement. A total optical depth was computed using the Langley method. The ozone optical depth and Rayleigh optical depth were calculated based on the TOC and altitude, respectively. Subtraction of these two components yields the AOD. For a clear day, it can be shown that the fractional uncertainty of total optical depth ranges from 0.1 percent at 300 nm to 1.9 percent at 368 nm, which demonstrates the stability of the retrieval method.
A pattern of higher AODs during the summer and lower AODs during the winter was observed for all wavelengths. For example, at 332 nm the summer mean AOD was 0.663 ± 0.26, the fall mean AOD was 0.217 ± 0.18, and the winter mean AOD was 0.076 ± 0.03. The higher optical depth values during the summer can in part be attributed to the persistent haze (of both anthropogenic and natural origin) that exists in the region throughout the summer. This haze significantly decreases the visibility in the nearby Smoky Mountain National Park and is a topic of great concern to the National Park Service. The large degree of variability in the AOD is due to the complex interaction of UV-B and aerosols. The aerosol optical properties include both scattering and absorption components that have a strong wavelength dependence, as well as a dependence on the physico-chemical properties of the aerosol.
The AOD increases as wavelength decreases. Ozone absorption becomes the dominant factor shaping the total optical depth at shorter wavelengths, whereas at the longer wavelengths, the aerosol and Rayleigh terms are the significant contributors. The increasingly strong absorption by ozone at the shorter wavelengths accounts for approximately 70 percent, 50 percent, and 30 percent of the total optical depth at 300, 305, and 311 nm, respectively. The Rayleigh optical depth increases from approximately 26 percent of the total optical depth at 300 nm to 70 percent at 368 nm. AOD also increasingly contributes to the total optical depth from approximately 7 percent at 300 nm up to 34 percent at 368 nm. However, it must be noted that the uncertainty in the AOD measurements at the shorter three wavelength channels is relatively large, and combined with the fact that ozone is the dominant attenuating factor at these wavelengths, decreases the confidence in the retrieved AODs at the lowest three wavelength channels.
The UVI is a forecast tool developed by the National Weather Service to inform the general public about the health hazards of exposure to UV radiation. The UVI is the predicted noontime erythemally-weighted UV irradiance in W/m2 converted to a unitless index scale. The Tropospheric Ultraviolet-Visible Radiation Model (TUV) is used to calculate UVI for cloud-free conditions at the field research site. Currently, the U.S. UVI uses a standard AOD input of 0.2, independent of wavelength, with no aerosol absorption (single scatter albedo, v = 1.0) for all 58 U.S. cities for which the UVI is generated. For each day that ozone and AOD were derived from the UVMFR over the 6-month period, the UVI was calculated using both the standard aerosol inputs and the actual retrieved AOD. Because the UVI standard AOD is referenced to 340 nm, Angstrom's formula was used for each day to extrapolate the measured spectral extinction to an AOD at 340 nm. The resulting difference between the two scenarios is computed as the UVI calculated using the measured AOD minus the UVI calculated using the standard AOD. Three scenarios of aerosol absorption are considered in conjunction with the measured AODs, no absorption (v = 1.0), moderate absorption (v = 0.90), and strong absorption (v = 0.75). The percentage decrease in erythemally-weighted UV for the three different absorption scenarios and each retrieved AOD was calculated. The percentage change is relative to the standard aerosol inputs (AOD = 0.2, v = 1.0). As the AOD increases, the decrease in surface erythemally weighted UV becomes substantial, up to 17 percent for the highest measured AOD for the no- absorption case. Inclusion of aerosol absorption in the calculation shifts to a larger decrease for each AOD. The extreme case of high absorption and high AOD indicated a nearly 50 percent decrease in surface erythemally weighted UV. The lesser aerosol attenuation during the low AOD cases can lead to an increase of up to 4 percent in the surface erythemally weighted UV.
Deviations in the UVI were calculated for 74 days, comparing the standard to the measured aerosol inputs. For the no-aerosol absorption scenario, the majority of the days (63.5 percent) had no change in the resulting UVI. However, there was a deviation of –1 index unit for 28.4 percent of the days where using the actual measured AOD results in a lower UVI prediction versus the standard aerosol inputs. There also were a few days (4.1 percent), where the UVI deviation was either –2 or +1 index units. Including a moderate amount of aerosol absorption increases the frequency of negative UVI deviations with no change for 59.5 percent of the days, –1 index unit was 20.3 percent, –2 index units was 13.5 percent, and –3 index units was 5.4 percent. Strong aerosol absorption shifted the UVI deviations further negative, with 44.6 percent of the days having no change, 23 percent were –1 index unit, 12.1 percent were –2 index units, 13.5 percent were –3 index units, 5.4 percent were –4 index units, and 1.4 percent were –5 index units. Grouping the deviations by month revealed a seasonal pattern, with a majority of the –2 and –1 UVI deviations occurring during the summer months of July and August and the late fall and winter months showing 0 or +1 deviations. The moderate and strong absorption cases showed similar seasonal patterns. Deviations grouped by measured AOD demonstrated that the higher the optical depth the greater the chance of a UVI deviation. These results indicate that the aerosol optical properties can have an impact on the UVI calculation, that during the summer months when AODs are typically higher, particularly in urban areas, the use of the standard AOD input of 0.2 can lead to an overestimation of the surface UV, and thus a higher UVI forecast value. In terms of improving the accuracy of the UVI forecast by including the effect of haze, it is suggested that during the summer months, a larger value of AOD and a more realistic value of aerosol absorption (v = 0.90) should be used. Based on the values retrieved in this study, an AOD in the range of 0.6 to 0.75 is recommended. The actual summertime AOD will naturally vary, dependent upon each forecast city location; however, it can be expected that in most urban locations, the AOD in the UV will be greater than the standard input of 0.2.
A procedure to retrieve single scattering albedo (
), the ratio of scattering coefficient to total extinction coefficient, in the
UV wavelengths has been devised. It is an important aerosol parameter in determining
the aerosol effect on UV radiation. The procedure involves coupling measurements
of surface UV irradiance from the UVMFR-Shadowband Radiometer (UVMFR-SR) with
the tropospheric ultraviolet radiative transfer model TUV4.1 from NCAR (http://www.acd.ucar.edu/TUV/ ).
The UVMFR-SR was deployed near the town of Black Mountain, NC (35.63N, 82.33W,
951 m amsl) atop a 10 m tower. TOC and AOD values from 1999 were used as inputs
to TUV4.1. Assumed values of asymmetry parameter (g) and ground albedo (ga)
of 0.70 and 0.04, respectively, also were inputs. TUV4.1 was modified to output
diffuse-to-direct ratio (DDR) for the seven wavelengths of the UVMFR-SR at solar
noon. The initial output of DDR from TUV4.1 was compared with UVMFR-SR measurements.
The DDR was derived from the UVMFR-SR measurements as the ratio between the
recorded diffuse and direct irradiances. The TUV4.1 was iterated by varying
until the output
matched that of the UVMFR-SR, ultimately yielding
for each wavelength of the UVMFR-SR for each day. Only days with cloudless conditions
at solar noon were considered. Additionally, back trajectory analysis was used
to determine if values of
could be correlated with air mass classifications as determined at the research
site. The 48-hour back trajectories were computed from the site using the HY-SPLIT
Model for each value of
to see which sector the air over the site came from (model obtained from web
address http://www.arl.noaa.gov/ready/hysplit4.html ).
Single scattering albedo
values were obtained for the seven wavelengths of the UVMFR-SR for 9 days, from
July 26 to October 5, 1999. The value of
ranged from 0.53-0.94 at 300 nm, 0.58-0.99 at 305.5 nm, 0.59-1.00 at 311.4 nm,
0.59-1.00 at 317.6 nm, 0.59-1.01 at 325.4 nm, 0.59-1.01 at 332.4 nm, and 0.55-1.03
at 368 nm. For a few of the days, the upper range of
for the longer wavelengths (325.4, 368 nm) is not physically possible. However,
if the uncertainty in the procedure used to obtain these values is factored
in, they can be said to be unity or very close to unity. There was no evidence
of correlation between air mass type and
at these wavelengths. This can be attributed to the widely varying aerosol content
of air masses coming from each sector. This is especially true for air masses
originating in the polluted sector, where efficient scatterers (sulfate) and
absorbers (soot) abound.
It was necessary to perform sensitivity studies so that error in this
retrieval technique could be determined. Tests were conducted to see which input
played the greatest role in determining DDR. One parameter was varied holding
the others constant. TUV4.1's output of DDR was expected to be more sensitive
to a more turbid atmosphere, so two AOD scenarios (0.3 and 0.8) were used. From
these tests the order of importance for parameters for determining DDR, based
on the rate of increase of DDR with respect to the variable in question, is:
AOD, single scattering albedo ,
asymmetry parameter (g), ga, and TOC. It was found that DDR has little dependence
to realistic values of TOC. This can be attributed to the good absorption and
poor scattering properties of ozone in the UV wavelengths. Variations of tropospheric
ozone do little to influence UV radiation when compared to stratospheric ozone.
It was found that the sensitivity of DDR to
increased from AOD = 0.3 to AOD = 0.8, and this also was found to occur for
g and ga. A further test was conducted to see how much a change in DDR of 0.02
would affect the retrieval of
as AOD is varied. Here it was found that,
as AOD decreases, the change in
increases. Hence, error analysis in
retrieval was conducted at AODs, from 0.05 to 1 in 0.05 increments to see how
AOD affects the error.
At each AOD, the uncertainty in DDR values due to uncertainty in g, ga, and
AOD was determined. This was accomplished by summing together the greatest possible
error in DDR due to these three parameters. This error in DDR was combined with
the uncertainty in the instrument's DDR measurements through the root mean square
error (RMSE) formula. After the total error in DDR was found, how much this
error in TUV4.1 output of DDR affected the
values retrieved could be determined. The value of
was allowed to vary from a value of 0.86. The following displays the assumed
values of the model parameters and the uncertainty assigned to these parameters
and the instrument measurements:
| Assumed Value | Error (+/-) | |
| uncertainty in AOD due to Vo | 0.01 | |
| asymmetry parameter | 0.7 | 0.05 |
| ground albedo | 0.04 | 0.02 |
| diffuse-to-direct ratio | 0.02 | |
| single scattering albedo | 0.86 |
The uncertainty in AOD should be greater than ± 0.01 because in the UV wavelengths
AOD is not the dominating attenuator and is subject to changes in the other
attenuators. Ozone absorption is most important in the two shortest wavelengths
of the UVMFR-SR, and Rayleigh scattering is most important in the five longest
wavelengths. AOD accounts for only 7 percent of the total optical depth at 300
nm, increasing to 34 percent at 368 nm. The error related to this phenomenon
has not yet been quantified, but is expected to give greater uncertainty to
AOD measurements as wavelength decreases. The results of the error/sensitivity
analysis are as follows. There is no correlation between wavelength and uncertainty
in . However, the
increasing uncertainty in AOD retrieval with decreasing wavelength in the UV
spectrum also is expected to increase uncertainty in
retrieval as wavelength decreases. Uncertainty in this technique seems to decay
exponentially with increasing AOD, from an average error of ± 0.19 for AOD =
0.5 to an average error of ± 0.2 for AOD = 1.0. For AOD < 0.3, the estimated
error is ± 0.04. This was deemed to be the threshold for a reasonable
retrieval, so results from days with AOD < 0.3 were excluded.
The values of found
in the current work encompass a wider range than the few previous studies of
single scattering albedo in the UV wavelengths. One possible cause of the wide
range of values of single scattering albedo in this study, as compared to previous
work, is the high temporal and spatial variability of tropospheric aerosols.
The vast differences in sources, transformation and removal processes, and lifetimes
all contribute to this variability. Through comparison of the values of
found here, to the values found at the site in 1995 at 312 nm (0.75 - 0.93),
it is possible that there has been a recent change in the aerosol composition
in the air masses influencing the site. An increase in the scattering ability
or decrease in the absorbing ability of the aerosols at the site would explain
higher values of ,
and vice versa for lower values.
It was found that, on average,
increases with increasing wavelength, from 0.75 at 300 nm to 0.83 at 368 nm.
Statistical Analysis System (SAS) analysis indicated that the increase is statistically
significant at the 0.05 level in these wavelengths. It has been previously suggested
that an assumption that BC absorption is independent of wavelength can force
to decrease with
wavelength. This may explain the increase in wavelength of the current study,
as TUV4.1 has no unique input for black carbon aerosol properties. Absorption
and scattering properties of soot aerosols are allowed to vary within the input
of single scattering albedo.
The values of found
here can be used for better estimation of the parameter in these wavelengths
for the Southeastern United States. This will lead to further development of
UV radiative transfer models and lessen errors in estimation of surface UV irradiances
for the region.
References:
Chylek P, Banic CM, Johnson B, Damiano PA, Isaac GA, Leaitch WR, Liu PSK, Boudala
FS, Winter B, Ngo D. Black carbon: atmospheric concentrations and cloud water
content measurements over southern Nova Scotia. Journal of Geophysical Research
1996;101(D22):29105-29110.
Jennings SG, McGovern FM, Cooke WF. Mass concentration of aerosol black carbon at Mace Head on the west coast of Ireland. Atmospheric Environment 1993;27A:1229-1239.
Lenoble J. Atmospheric Radiative Transfer. A. Deepak, Virginia, 1993, 532 pp.
Journal Articles on this Report: 7 Displayed | Download in RIS Format
| Other project views: | All 34 publications | 17 publications in selected types | All 15 journal articles |
| Type | Citation | ||
|---|---|---|---|
|
|
Bahrmann CP, Saxena VK. Influence of air mass history on black carbon concentrations and regional climate forcing in southeastern United States. Journal of Geophysical Research - Atmospheres 1998;103(D18):23153-23161. |
R825248 (Final) |
not available |
|
|
Frederick JE, Slusser JR, Bigelow DS. Annual and interannual behavior of solar ultraviolet irradiance revealed by broadband measurements. Photochemistry and Photobiology 2000;72:488-496. |
R825248 (Final) |
not available |
|
|
Im JS, Saxena VK, Wenny BN. Temporal trends of black carbon concentrations and regional climate forcing in the southeastern United States. Atmospheric Environment 2001;35(19):3293-3302. |
R825248 (Final) |
not available |
|
|
Modrak MT, Saxena VK. Contributing parameters to visibility degradation in western North Carolina. Atmospheric Environment, Washington, DC, 2000. |
R825248 (Final) |
not available |
|
|
Petters JL, Saxena VK, Wenny BN, Madronich S. Aerosol single scattering albedo retrieved from measurements of surface UV irradiance and a radiative transfer model. Journal of Geophysical Research-Atmospheres 2003;108(D9):Art.No. 4288. |
R825248 (Final) |
not available |
|
|
Wenny BN, Saxena VK, Frederick JE. Aerosol optical depth measurements and their impact on surface levels of ultraviolet-B radiation. Journal of Geophysical Research - Atmospheres 2001;106(D15):17311-17319. |
R825248 (2000) R825248 (2001) R825248 (Final) |
not available |
|
|
Yu S, Saxena VK, Wenny BN, DeLuisi JJ, Yue GK, Petropavlovskikh IV. A study of the aerosol radiative properties needed to compute direct aerosol forcing in the Southeastern United States. Journal of Geophysical Research 2000;105(D20):24739-24749. |
R825248 (Final) |
not available |
atmosphere, stratospheric ozone, UV-B radiation, solar radiation, modeling, southeast United States. , Air, Scientific Discipline, RFA, climate change, Atmospheric Sciences, Environmental Chemistry, tropospheric ozone, UV-B radiation, Global Climate Change, urban air, ozone, weather factors, ambient aerosol particles, climate variations, aerosol sampling, environmental monitoring, boundry layer processes, climate variability, ambient ozone data, air quality data, air pollution models, atmospheric monitoring
Relevant Websites:
http://www4.ncsu.edu/unity/users/s/saxena/public/cloud.html
Progress and Final Reports:
1997 Progress Report
1998 Progress Report
1999 Progress Report
2000 Progress Report
2001 Progress Report
Original Abstract