Previous: Introduction Next: Results and Discussion Up: Ext. Abst.
Instrumentation and methodology
To measure UV irradiance the Brewer uses a quartz dome and Teflon diffuser with a hemispherical field of view. For the EPA/UGA network, a dynamic schedule is used by the Brewer allowing UV readings to be recorded approximately every 30 minutes throughout the day, ensuring that a UV scan coincides with solar noon. The Brewer has a UV spectral range of 286.5 to 363 nm in 0.5 nm steps. UV irradiance calibrations, using a secondary standard lamp traceable to a NIST 1000 W lamp, are performed at the sites by NUVMC. Resulting response functions were used to calculate irradiance from photon counts. Calibrations are targeted to occur once per year. In addition independent quality assurance audits of the instruments take place.
The Brewer recorded total column ozone levels between UV measurements. The methodology of Sabburg et al. (2000) was used to obtain direct sun ozone values to less than ± 3 %.
The data presented in this paper have been corrected for dark count, dead time, stray light, UVA, cosine response, temperature dependence and temporal response. Only limited quality assurance (QA) has been performed on the raw data, such as removing extreme outliers. The uncorrected data includes dark count, dead time, UVA correction and uses the latest response function.
Sabburg and Meltzer (2000) explain the cosine correction methodology associated with the full sky diffuser in detail. In summary, cosine response measurements were made on each Brewer using the irradiance of a standard 1000 W lamp. These measurements were performed in the laboratory. The final values were based on an average of measurements along the long and short sides of the Brewer, five wavelengths and two sets of zenith angles. The wavelengths were 306.3, 310.1, 313.5, 316.8 and 320.1 nm. The lamp rotated over the zenith angle range of ö 80 to 0 ° and 0 ° to 80 ° in 10 ° steps.
The equations of Bais et al. (1998) were used to calculate the total cosine correction assuming a diffuse isotropic sky and the ratio of the direct / global irradiance was based on the clear-sky model of Rundel (1986). The model used corrected ozone amounts from the Brewer and a typical value of aerosol optical depth (AOD) was chosen for each site, namely, 0.2, 0.1, 0.8 and 0.5 for Boulder, Rocky, Gaithersburg and RTP respectively (Jeral Estupinan, Personal Communication, 2000). When ozone data was not available, the nominal value of 300 DU was used for that day.
The temperature dependence of two of the Brewers (#101 and #146), was measured by the NUVMC during a field campaign in July 2000. The methodology for measuring the temperature dependence is outlined by Meltzer et al. (2000). The Brewer temperature fluctuates with the ambient at the various locations from 0 to + 50 ° C. The temperature dependence from - 18 to + 42 ° C of the response of three MKIV Brewers (not the Brewers in this current paper), has been determined in the laboratory using a controlled variable temperature environment. The observed temperature coefficients of their responses at 306 nm were - 0.17 %, - 0.22 % and - 0.37 % per ° C. This results in a predicted seasonal variation of their sensitivity of up to approximately 20 %, well beyond the desired accuracy of ± 3 %. There is a significant wavelength dependence of the temperature coefficient below 325 nm. This is primarily due to the temperature dependence of the transmission of a nickel sulfate filter.
The methodology also requires that local operators obtain the temperature coefficients of the response as a function of wavelength at each site. The method utilizes spectra of the 50 W Brewer calibration lamps recorded throughout the day during the diurnal temperature cycle. These measurements require more accurate and stable instrumentation than is supplied with the Brewer. Plots of the photon counts versus temperature at each wavelength were used to determine a temperature coefficient, D R/D T, which is the slope of the response versus temperature. The temperature corrections were achieved by normalizing photon counts to an equivalent photomultiplier tube (PMT) temperature of 20 ° C.
The temporal variation in the instrument response is due to optical changes in the characteristics of the instruments. This necessitates an annual UV irradiance calibration, using a secondary standard lamp traceable to the NIST 1000 W lamp, to be performed at the sites by NUVMC staff. Resulting response functions are used to calculate irradiance from photon counts. In addition, NOAA, using similar equipment to the NUVMC, conduct independent quality assurance audits of the instruments. Details of these procedures are available at the UGA web address: ftp://oz.physast.uga.edu/Outgoing/ by downloading the three documents entitled: SOP1_FEL-Lamp.doc, SOP for Field Calibration.doc and Irradiance Transfer of 1000 Watt lamps.doc.
C) Model UV data
In order to make a comparison of the corrected UV data (in addition to comparing to the uncorrected data), the data calculated by the clear sky model, Uvspec, was used (Kylling 1995). UVSPEC used the wavelength range of 280 to 400 nm with 1.0 nm resolution and used two-stream code to solve the radiative transfer equation. Ozone and oxygen absorption, as well as Rayleigh scattering was accounted for. UVSPEC used the same nominal AOD values as used for the cosine correction, scaled according to visibility. When ozone data was not available, the nominal value of 300 DU was used for that day.
D) Satellite Data
Finally, a comparison is made with the inferred DUV data from the NASA, Total Ozone Mapping Spectrometer (TOMS). TOMS data is indicative at the time of over-pass (approximately 11:15 AM local time) and the footprint for the TOMS instrument is approximately 40 km x 40 km. The TOMS erythemal exposures are calculated with a radiative transfer model (Herman et al. 1996; Udelhofen et al. 1999). The TOMS algorithm uses ozone data as well as reflectivity measurements at 380 nm to identify cloudy scenes. The daily integration is carried out assuming no diurnal variation in cloudiness.