2008 Progress Report: Effects of Clouds and Tropospheric Air Quality on Surface UV at 6 UV Research SitesEPA Grant Number: R833224
Title: Effects of Clouds and Tropospheric Air Quality on Surface UV at 6 UV Research Sites
Investigators: Lantz, Kathleen O. , Kiedron, Peter , Petropavlovskikh, Irina
Institution: Cooperative Institute for Research in Environmental Sciences
EPA Project Officer: Chung, Serena
Project Period: October 1, 2006 through September 30, 2010
Project Period Covered by this Report: October 1, 2007 through September 30,2008
Project Amount: $299,988
RFA: Implications of Tropospheric Air Pollution for Surface UV Exposures (2005) RFA Text | Recipients Lists
Research Category: Global Climate Change , Air Quality and Air Toxics , Climate Change , Air
Six locations across the continental United States will be used to evaluate the impacts of tropospheric air quality on surface UV irradiance measurements. These six sites are the re-established EPA UV Network and include Table Mountain near Boulder, CO, Rocky Mountain Research Station at Niwot Ridge, CO, Bondville, IL, Fort Peck, MT, Raleigh, NC, and Houston, TX. Our goals will be to evaluate UV-B irradiance from the six locations under a variety of atmospheric conditions.
The specific goals include (1) developing and implementing QA/QC procedures on the UV measurements and ancillary data. Procedures include comparing solar irradiance from the Brewer spectrophotometer to measurements from other co-located instruments (i.e. UV broadband radiometers and UV filter radiometers). (2) An algorithm will be developed for determining ozone profiles from Brewer Umkehr measurements and inferring tropospheric ozone column. (3) Cloud and aerosol properties will be collected from a suite of co-located instruments as part of the SURFRAD Network within the Surface Radiation Research Branch (SRRB), Earth System Research Laboratory, NOAA, and the USDA UV monitoring program. (4) Using the direct-to-diffuse ratio from the UV-MFRSR combined with a UV radiative transfer code, the aerosol single scattering albedo will be estimated for clear-sky conditions at the six sites. (5) The atmospheric conditions will be characterized for total ozone, cloud properties, and atmospheric pollutants. (6) We will provide a comprehensive evaluation of the impact of total ozone, and air quality on the tropospheric UV-B transmission. The database provided by NEUBrew, the Central UV Calibration Facility, the Surface Radiation and Research Branch of ESRL/NOAA, and the USDA UV Monitoring Program provides a unique and valuable data-set that will help identify the impacts of variable conditions of tropospheric air quality and cloud cover on UV irradiances.
This document describes results and progress during the second year under the objectives of the EPA STAR funded proposal “Effects of Clouds and Tropospheric Pollution on Surface UV at six EPA UV Research Sites”. The project’s primary focus is the study of changes in surface UV levels caused by atmospheric pollutants, including aerosols and tropospheric ozone. The re-established NOAA/EPA UV Network at six locations across the United States is currently measuring UV radiation and automated direct sun and zenith sky ozone measurements from which daily ozone profiles are derived.
Database and QA/QC procedures and development: Procedures were developed and implemented for generating higher-level spectral solar irradiance (level 201) both as web generated graphs and ftp ascii files. Both files and plots are available from the NEUBrew web-site, http://www.esrl.noaa.gov/gmd/grad/neubrew Exit . The corrections include spike detection and spike removal algorithms [Kiedron et al., May 2008] and stray light correction [Kiedron et al., June, 2008]. Level 201 spectral solar irradiance is used to generate UV dose and erythema according to algorithms outlined in Kiedron et al., December 2007. General information on NEUBrew and new products was presented at the Quadrennial Ozone Symposium, Tromso, Norway [Disterhoft et al., 2008].
In addition to erythema and UV index from the Brewer spectrophotometer at each of the sites, erythema and UV index can be derived from the co-located UV broadband radiometer. Solar zenith angle dependent erythema calibration factors were applied to the UV broadband radiometers to calculate an erythemally weighted solar irradiance. The comparison between the two instruments has been made available on the NEUBrew web-site (http://www.esrl.noaa.gov/gmd/grad/neubrew/ Exit . The accuracy of the calibration of the broadband radiometers for erythema used for this work was verified in an intercomparison between international UV calibration facilities [Hulsen et al., 2008]. Results indicated a difference of 0.2% at high sun from the Physikalisch-Meteorologisches Observatorium (PMOD) calibration reference in Davos, Switzerland. Previous work has shown that YES UVB radiometers used in the SURFRAD and USDA UV monitoring Network have an annual stability or variation on average of approximately 2% [Lantz et al., 2005]. Future work will check the changes in UV responsivity of the Brewer spectrophotometer with the YES UVB radiometers at a given solar zenith angle.
Ozone column: Ozone column data is generated using internal Brewer spectrophotometer procedures from direct sun (DS) observations. Average daily total ozone values are obtained from measurements with air masses less than four and for cases when the standard deviation of five individual measurements constituting the DS procedure is less than 2.5 DU. From here, level 200 ozone column data are generated and are corrected for the throughput instability. Level 200 data is implemented because the effective extraterrestrial constant is affected by instrument throughput -where in the case of the MKIV Brewer spectrophotometer is chiefly sensitive to the solar blind filter that is made of hygroscopic NiSO4 crystal. The correction uses two or more daily internal quartz tungsten halogen lamp measurements. The level 200 ozone column significantly improved correlation with OMI data for the few Brewers that experienced drift due to solar blind filter instability. In particular we obtained encouraging results with Brewer #139 that had the NiSO4 filter altogether replaced with Schott UG11 color glass filter. This result gives us confidence that the expensive NiSO4 filter is not always necessary, and in the future it might be replaced with a UG11 filter. The level 200 data ozone column data are available as web graphs and the ftp ascii text files will be released shortly. Extended data sets will be available for future satellite mission validation and ozone recovery analysis.
Stray light errors in the ozone column data from MKIV Brewer spectrophotometers were evaluated and a limit for maximal slant path column was established to be about 1200 DU [Kiedron et al., June, 2008]. The possibility of a different mathematical algorithm for ozone retrieval was investigated that could partially correct the stray light error. This algorithm has not been implemented in routine calculations.
Ozone retrievals are dependent on ozone profiles (particularly for large air masses) and on the effective profile temperature. Using UV-RSS Table Mountain data we explored the possibility of detecting the effective ozone profile temperature. These results were presented at the Ozone symposium [Kiedron and Michalsky, Ozone Symposium, 2008].
Ozone profiles and tropospheric ozone: The new Brewer ozone algorithm software called O3BUmkehr was developed in collaboration with Martin Stanek of the International Ozone Service, Canada. The Brewer Umkehr ozone profile retrieval algorithm is implemented at all NEUBrew sites to derive daily ozone profiles. The work to further optimize the Brewer ozone profile retrieval has continued through 2008. For example, the option was added to the processing of the Brewer data to adjust total ozone information for drifts in the standard lamp readings based on the work discussed above for total ozone by Kiedron and colleagues. For this task, the standard lamp (SL) is placed at the beginning of the optical part in the Brewer. The double ratios R5 and R6 are calculated from sunlight during direct-sun (DS) total ozone measurements. It is assumed that these ratios are stable. If the SL is stable, all changes of the calculated R5/R6 from this lamp reflect the changes in the optics or photomultiplier. Then, the change in R6 ratio can be used to correct the extra-terrestrial (ETC) constant. Another implementation to the software was a “Raw data” panel to evaluate cloudy periods on a daily basis [Flynn et al., AGU, Fall 2008].
This work evaluated the quality of tropospheric ozone information derived from the ground-based Brewer zenith sky measurements. Monitoring of the day-to-day and diurnal tropospheric ozone changes is one of the science objectives of this grant. Topospheric ozone data was evaluated through comparisons with co-incident ozonesonde measurements of high vertical resolution. Since the Umkehr method provides vertical profile information that is smoothed over an altitude range, the Umkehr smoothing functions (or Averaging Kernels) can be used for comparisons with other high-resolution ozone profile data sets. Therefore, the Averaging Kernel (AK) coefficients are saved in a separate file for 16 Umkehr layers (top to bottom). The file also includes the 16-layer a-priori ozone profile. This information is used for calculations of the smoothed ozone sonde profiles.
The analysis concentrated on the short-term and long-term tropospheric ozone variability detected by co-incident and co-located Brewer data along with ozone profiles from ozone-sondes available from Boulder, CO. Total of five Brewer instruments located in Boulder, CO were used in the analysis. Some instruments were not calibrated for total ozone. In this case a significant bias was found between the Brewer estimates of the tropospheric ozone column (Umkehr layer 1) and integrated ozone-sonde profile below 250 mb. The studies of the day-to-day variability in the tropospheric ozone suggest that no more than 1-day separation is allowed between the ozone-sonde launch and Brewer Umkehr measurement. It appears to be especially important in the winter and spring time period of measurements in Boulder area, when there is a high frequency of high-latitude air-mass intrusions in the middle latitudes. Never-the-less, analyses suggest that Umkehr technique performed by the well-calibrated Brewer is capable of monitoring short-term variability in tropospheric ozone. It can explain about 50 % of the variability measured by ozone sonde [Petropavlovskikh et al., Ozone Symposium, 2008].
Stray-Light and Polarization: The work to optimize the Brewer algorithm for scattered light and polarization instrumental effects was performed in collaboration with P. Disterhoft (NOAA/ESRL/GMD) and A. Cede (NASA/Goddard). Intensive work has been done in the past to calibrate and maintain Umkehr instruments for stability and accuracy of the measurements [Komhyr, 1980, Komhyr et al. 1989; McElroy et al, 1995; WMO, 1999; Cede et al., 2006; McElroy, 2007]. Zenith-sky measurements have relied on the total ozone instrument calibration and the normalization procedure. Up until now, internal-scattering effects (stray light) on the zenith-sky measurements have not been well quantified. The effects of stray light can become significant when the sun is low and sensitivity of the instrument is reduced. These effects can produce errors in ozone profile retrievals and are partially responsible for traditional Dobson Umkehr retrievals being as much as 10% lower than corresponding SBUV ozone [Petropavlovskikh et al., 2005; 2008]. Although, Brewer MKIV instruments (employed by the former EPA UV network) have scattered light rejection comparable to the Dobson instrument (~10-5), the single Brewer instruments have not been fully characterized for the effects of the stray light on ozone profile retrievals. Due to its optical design, the Brewer Mark III instrument deployed at NASA/Goddard has a very low level of the stray light (10-7) that contributes to the zenith sky measurements.
The spectral transmission of the double and single Brewer spectrophotometer is defined by the transmission characteristics of filters, grating, PMT, etc. The work to assess the effects of the stray light in the Brewer Umkehr retrievals were performed in Boulder on September 20 and 29, 2007. Five of the NEUBrew instruments employed in Boulder were used in the assessment. The synchronized Umkehr measurements were also taken by four Dobson instruments during sun rise between 60 and 90 degrees SZA. The ozone profiles were retrieved for each Umkehr curve and compared to the reference ozone profile compiled from ozone sounding and satellite data. The coincident launches of ozonesondes were performed at the site near Boulder. The Aura Microwave Lidar Sounding (MLS) V2.2 satellite ozone profiles were selected by the over-pass location closest to Boulder, and are used to extend the ozone sounding above the burst level of ozonesondes [Petropavlovskikh et al. 2008]. MLS data were made available through the NASA AVD web page (http://avdc.gsfc.nasa.gov/ Exit ) [Petropavlovskikh et al., 2008; Petropavlovskikh et al., Ozone Symposium, 2008].
The simulation of the Umkehrs for Brewer C-pair wavelengths was done through convolution of the zenith sky radiances with the core band-passes only. The stray light contribution weighting was done with the 2e-5 coefficient, which seems to have the best fit for the case. The correction method for the stray light contribution is not perfect and appears to vary for each instrument. It appears that the stray light corrected N-values tend to slightly over or underestimate the reference at some SZA and the “spoon” shape is not completely removed. However, the maximum of the stray light contribution to the Umkehr curve is found close to 85 degrees SZA.
The corresponding errors in the retrieved ozone profiles relative to the combined ozone sounding and MLS profiles vary with altitude. The retrieved ozone shows about 5 to 10 % underestimation in the stratosphere (layer 7, 8 and 9), and about 5-10 percent overestimation at the ozone maximum (layer 4), while in the troposphere (layer 1) some instruments show about 5-10 % lower ozone (Brewer 141 and 134), and some instruments show very good agreement (Brewer 105). The worst agreement is found for the instrument 105, which was not calibrated for total ozone. Including the stray light correction seems to reduce the retrieval error, however the generalized estimation method used in these analysis does not completely remove it. The remaining errors suggest that different instruments have different levels of the stray-light throughput at wavelengths outside of the core band-pass. Therefore, more detailed investigation of the stray light contribution is needed for each instrument. The corresponding errors in the retrieved ozone are vertically distributed with lower/higher ozone values retrieved in the stratosphere/troposphere, while the total ozone is preserved. This distribution of the offset in the upper and lower part of the ozone profile is similar to the previous findings of biases between Umkehr and other types of ozone measurements.
Comparisons of Brewer UV Index to NWS UV Index Forecasts: The NEUBREW network (NOAA-EPA UV Network) was designed to investigate factors affecting surface UV solar irradiance. UV index from the Brewer spectrophotometers for the six sites are compared to the NCEP UV forecast and plots are available next day on the NEUBrew web-site. Early results indicate that at the “dirtier” sites the UV forecast is high with respect to the measurements. However, this needs to be re-evaluated after further QA/QC levels of the Brewer data have been generated. The UV forecast considers forecasted ozone and clouds, ground albedo, surface elevation, and aerosol properties for the calculations [C. Long et al., 1996]. The discrepancies set the stage for evaluating measured parameters representing air quality and clouds on solar uv radation at the sites.
UV Surface Albedo: Accurate measurements of the surface albedo are necessary for the retrievals of aerosol single scattering albedo and for studying factors affecting surface UV solar irradiance. In March 2008, a tower was installed at the Table Mountain Test Facility to measure surface albedo in the UV and the visible. The tower is approximately 25 feet tall with two 3 feet extensions perpendicular to the tower that hold the downward viewing vis-MFRSR and a broadband radiometer for measuring up-welling solar radiation. On the TMTF deck there is a vis-MFRSR and a UV broadband radiometer to measure downwelling solar radiation (i.e. 415 – 868 nm and erthemally-weighted solar irradiance, respectively). The four instruments were calibrated prior to installing the tower with the radiometers. Calibrations of the UV instruments are scheduled once per year. Previous studies have shown the UV broadband radiometers to be very stable with respect to each other, which makes this an ideal instrument for measuring surface albedo [Hulsen et al., 2008; Lantz et al., 2005]. In addition, spectral UV surface albedo measurements were made with a UV-MFRSR and a UV broadband radiometer on a short-term basis with the up-welling measuring instruments maintained at an 8 ft height. The short-term measurements were compared to the tower measurements and were in good agreement for this site. The spectral UV surface albedo measurement gave a surface albedo of 1.7% to 3.1% from 305-nm to 368-nm.
Aerosol properties and algorithm development: The first successful deployment of the fully- operational ultraviolet rotating shadow-band spectroradiometer occurred during the May 2003 US Department of Energy’s Atmospheric Radiation Measurement program’s Aerosol Intensive Observation Period. The Atmospheric Radiation Program (ARM) site is located in northern Oklahoma. The aerosol properties in the visible range were characterized using redundant measurements with several instruments to determine the column aerosol optical depth, the single scattering albedo, and the asymmetry parameter needed as input for radiative transfer calculations of the downwelling direct normal and diffuse horizontal solar irradiance in clear-sky conditions. The Tropospheric Ultraviolet and Visible (TUV) radiative transfer model developed by Madronich and his colleagues at the US National Center for Atmospheric Research was used for the calculations of the spectral irradiance between 300–360 nm. Since there are few ultraviolet measurements of aerosol properties, most of the input aerosol data for the radiative transfer model are based on the assumption that UV input parameters can be extrapolated from the visible portion of the spectrum. Disagreements among available extraterrestrial spectra, which are discussed briefly, suggested that instead of comparing irradiances, measured and modeled spectral transmittances between 300–360 nm should be compared for the seven cases studied. Transmittance was calculated by taking the ratios of the measured irradiances to the Langley-derived, top-of-the-atmosphere irradiances. The cases studied included low to moderate aerosol loads and low to high solar-zenith angles. A procedure for retrieving single scattering albedo in the ultraviolet based on the comparisons of direct and diffuse transmittance is outlined. The strength of this method is in using transmittances to avoid uncertainties in the extraterrestrial fluxes and by first fitting the direct and diffuse transmittance separately, the accuracy of the aerosol optical depth can be verified. For this study, five days in the spring of 2003 were used to evaluate the method. The retrieved UV single scattering albedo for this location are typically lower by 0.03 to 0.1 from the SSA at 550 nm, except in one case where the SSA was higher by 0.015. Specifics of the results of this work are described in Michalsky and Kiedron .
For the NEUBrew network, single scattering albedo was calculated for the Table Mountain test facility site (TMTF). Each of the six NEUBrew sites has a UV Multi Filter Shadowband Radiometer (UV-MFRSR), but the TMTF site also has a collocated UV rotating shadowband spectrograph (UV-RSS) that provides diffuse and direct spectral UV solar irradiance from 300 – 360 nm with a spectral resolution that varies between 0.25 and 0.45 nm. Inputs to the SSA retrieval algorithm included aerosol properties in the UV and/or extrapolated from visible measurements (i.e. aerosol optical depth, Angstrom coefficient, asymmetry parameter), surface albedo, total ozone, and ozone profiles. Information was obtained from instrumentation and products from NEUBREW, SURFRAD, USDA UV monitoring program, and AERONET. Direct-to-diffuse solar irradiance ratios (DDR) from a co-located UV Rotating Shadowband Spectrograph (UV-RSS) and a UV-MFRSR have been compared further as part of the development of an algorithm for the calculation of aerosol single scattering albedo (ω0). Initial results indicate good agreement between the two instruments at high sun for the channels from 305 – 368 nm. However, discrepancies in DDR at low sun (70°) are as much as 10% for specific channels. The discrepancies appear to follow the angular response errors of the UV-MFRSR channels and indicate that the angular response corrections to the signals may need to be re-evaluated an improved. A change in DDR of 0.05 gives an error of approximately 0.06 in ω0 when the aerosol optical depth is 0.2. Results gave a single scattering albedo from 0.84 to 0.91 for the UV spectral range. Results were presented at the American Geophysical Union, Fall Meeting, 2008 [Lantz et al., AGU, Fall 2008].
Michalsky and Kiedron  and Lantz et al. [December, 2008] showed that radiative transfer models may lead to retrieval of aerosol single scattering albedo (SSA). However to perform the retrieval successfully an accurate knowledge of aerosol optical depth (AOD) is necessary. AOD retrievals using UV-RSS data were presented at an ARM Meeting [Kiedron et al., November, 2008]; where the AOD from the UV-RSS will be used to verify the quality of AOD retrieval algorithm from the Brewer data. The algorithms for AOD retrieval with Brewer spectrophotometer data are under development. The issue of ozone cross-section, i.e., its uncertainty, was also evaluated.
Network expansion: The Atmospheric Radiation Measurements (ARM) program of the Department of Energy is interested in expanding their measurements to UV solar radiation. We have proposed to provide and run Brewer spectrophotometers at their sites [Kiedron et al., November, 2008]. Currently the deployment at the Southern Great Plains site in Oklahoma is slated for 2010.
QA/QC: For UV solar irradiance measurements, we will correct the angular response error in the Brewer spectrophotometer data using the diffuse and direct solar irradiance from the UV-MFRSR radiometers. We will check and potentially correct the stability of the UV solar irradiance from the Brewer using co-located UVB broadband radiometer data.
For total ozone measurements, we plan to develop a Langley regression based method to monitor the extraterrestrial constant that would be independent of measurements with the internal lamp. This method, if successful, could effectively eliminate or reduce the frequency for external Brewer calibrations such as the one performed by Environment Canada.
For ozone profile retrievals, we will continue analysis of the stray light contribution by simulating Brewer Umkehr measurements in Brewer MKIII and MKIV instruments. We will utilize the measured filter transmissions for individual NEUBrew instruments (by P. Disterhoft, NOAA/ESRL) and for Goddard MKIII instrument (by A. Cede, NASA,/Goddard), and will implement the measured slit functions for Brewer instruments in simulating the Umkehr sky radiances. This approach will help with optimization of Brewer ozone profile retrievals.
Tropospheric ozone column: We will also continue to build the reference data set in Boulder by using co-located ozone-sonde measurements launched in Boulder, CO on weekly bases to test limitations of the Umkehr derived tropospheric ozone variability. In addition, we plan to utilize the Boulder County, CO network of the surface ozone meters situated along the Boulder canyon. The instruments are spaced at about 200-300 m intervals along the 2000 m elevation gradient. The network is operated by NOAA/ESRL and University of Colorado. The Co-I S. Oltmans will be responsible for the data processing and formatting (at no cost to this proposal). Preliminary comparisons of the ozone altitude gradient in surface ozone data against the ozone-sounding indicate that surface ozone measurements can be used in place of ozone-sounding for vertical ozone distribution studies [results presented at the NOAA/ESRL 2008 Global Monitoring Annual Conference]. The advantage of surface ozone measurements is that they provide continuous data (hourly averages), whereas ozone-sounding is launched once a week.
UV radiation and air quality: Methods that were developed last year for retrieving aerosol single scattering albedo will be applied to more cases at the Table Mountain site and to more polluted sites where necessary ancillary information are available. The TMTF is not the ideal site given the aerosol optical depths have been found to be very low and at the limit of the retrieval uncertainties. The algorithms for AOD retrieval with Brewer spectrophotometer data will be further developed. In the final year of the project, the effect of tropospheric ozone and air quality on the surface UV solar irradiance will be accessed after the aerosol properties have been determined.
Finally, two NOAA Hollings scholarship recipients will be mentored in the summer of 2009 to study the effects of air quality and clouds on UV, and to compare ozonesondes, with Brewer ozone profiles.
K. Lantz, P. Disterhoft, C. Wilson, G. Janson, J. Slusser, Out of Band Rejection Studies of the UV multi-filter rotating shadow-band radiometers, Proceedings of SPIE, Remote Sensing of Clouds and the Atmosphere, Vol. 5979, 59791N-1, 2005.
Cede, A., S. Kazadzis, M. Kowalewski, A. Bais, N. Kouremeti, M. Blumthaler, and J. Herman (2006), Correction of direct irradiance measurements of Brewer spectrophotometers due to the effect of internal polarization, Geophys. Res. Lett., 33, L02806, doi:10.1029/2005GL024860.
Komhyr, W.D. (1980), Operations Handbook-Ozone Observations with a Dobson Spectrophotometer, WMO Global Ozone Research and Monitoring Project, Report No. 6.
Komhyr, W.D., R.D. Grass and R.K. Leonard (1989), Dobson Spectrophotometer 83: A Standard for Total Ozone Measurements, 1962-1987, J. Geophys. Res., D7, 94:9847-9861, July 20.
Long, C., A. J. Miller, H- T. Lee, J. Wild, R. Przywarty, D. Hufford, UV Index forecasts issued by the National Weather Servce, Bull. Meteor. Soc., 77, 4, 729 – 748, 1996.
McElroy, C. T. (1995), A Spectroradiometer for the Measurement of Direct and Scattered Solar Spectral Irradiance from On-Board the NASA ER-2 High-Altitude Research Aircraft, Geophys. Res. Lett., 22, 1361.
McElroy, C. T., 2007, The Tenth biennial WMO consultation on Brewer ozone and UV spectrophotometer operation, calibration and data reporting, Northwich, United Kingdom, 4 - 8 June 2007, World Meteorological Organization Global Atmosphere Watch, GAW Report No 176, Geneva.
Petropavlovskikh, I., Froidevaux, L., Shetter, R., Hall, S., Ullmann, K., Bhartia, P. K., Kroon, M., Levelt, P., J Geophys. Res., 113, D16, D16S41, 2008.
Petropavlovskikh, I., C. Ahn, P. K. Bhartia, and L. E. Flynn (2005), Comparison and covalidation of ozone anomalies and variability observed in SBUV(/2) and Umkehr northern midlatitude ozone profile estimates, Geophys. Res. Lett., 32, L06805, doi:10.1029/2004GL02, 2002.
World Meteorological Organization (WMO) (1999), Scientific assessment of ozone depletion: 1998, Report No. 44, WMO Global Ozone Research and Monitoring Project, Geneva, Switzerland.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
|Other project views:||All 35 publications||13 publications in selected types||All 6 journal articles|
||Hulsen G, Grobner J, Bais A, Blumthaler M, Disterhoft P, Johnsen B, Lantz KO, Meleti C, Schreder J, Vilaplana Guerrero JM, Ylianttila L. Intercomparison of erythemal broadband radiometers calibrated by seven UV calibration facilities in Europe and the USA. Atmospheric Chemistry and Physics 2008;8(16):4865-4875.||
||Michalsky JJ, Kiedron PW. Comparison of UV-RSS spectral measurements and TUV model runs for clear skies for the May 2003 ARM aerosol intensive observation period. Atmospheric Chemistry and Physics 2008;8(6):1813-1821.||