Spectral UV observations at the Royal Meteorological Institute in Uccle (Belgium) Hugo De Backer, Anne Cheymol, René Lemoine and Andy Delcloo Royal Meteorological Institute of Belgium, Uccle, Belgium Abstract. UV spectral measurements at the RMI another standard during the Brewerworkshop in (Uccle, Belgium) date back as far as 1989. Over the Halikidiki in 1998 [WMO, 1998]. Some lamps were long time period 1989-2004 the frequency of damaged during the transport. New lamps with observations and calibration checks has changed. calibration data were acquired the later that year. It A description of the software tools developed to was noticed that the calibration of the new lamps was manage the different calibration levels and the re- consistent with the calibration level in use in analysis of the whole set will be explained. The Thessaloniki. The calibration files of the old lamps methodology takes the temperature of the instrument were adjusted to match the same level. In 2003 the into account, and in this way the estimated error on calibration was checked with two 1000 W NIST the measurements is less than 5%. This reprocessed traceable lamps during a calibration and maintenance dataseries of spectral data is used to construct a data visit. It turned out that there was no need for change set of UV indices. The resulting time series will be of the calibration level. Therefore all calibration data examined for the temporal evolution, in relation with was generated assuming the corresponding the concurrent ozone observations and other ancillary calibration for the 50W lamps as they were data at the same site. established in 1998. Brewer #178 is a double monochromatior (MK III) and its spectral range extends from 286 to 363 nm. Introduction The calibration level was established with the same 50W lamps as Brewer #016. It was also checked and UVB intensities received at the Earth’s surface are confirmed with the 1000W-lamp system during the closely related with the ozone content of the calibration in 2003. atmosphere [Bais et al, 1993]. Although it is not the change in ozone in the stratosphere, but the possible Data reduction change in UVB radiation at the surface that has impact on biosystems, there are much more long-term The calibration of both instruments is established data sets on ozone than on UVB. in almost the same way. For each observation (scan Also in Uccle (near Brussels in Belgium, 50º48'N, of global irradiation, or scan of a calibration lamp, the 4º21'E, 100m) we have total ozone observations since following calculations are performed. The raw photon 1971 with a Dobson spectrophotometer. In 1983 counts are first corrected for the dark count and the Brewer #016 was installed. Initially it was only used dead time of the instruments. Then they are converted for total ozone observations. In 1989 the instrument to counts per second. Finally the temperature was automated and equipped with the UVB dome. correction is applied. Since that time we started the UVB measurements. In Further processing depends on the purpose of the 2001 Brewer 178 was installed, and since that time observation. both Brewers operate side-by side. This offers the If a calibration with a lamp is performed (ul or xl opportunity to compare the reliability of both routine for the single and the double monochromator instruments. We can also have a look at the long-term respectively), the corresponding lamp spectrum is evolution of the UV intensities at the site. Before we read. Normally this is done with a set of scans on can do so we have to take care of the changes in the different lamps. From these data a fifth order function observation and calibration practices during this of the wavelength is fitted by least square regression period. to obtain a wavelength dependent response curve. For the double monochromator Brewer, this is done Instrumentation and calibration twice, once for the wavelength range 286-350nm (slit 1) and once for the wavelength range 350-363nm (slit As mentioned in the introduction we have two 5). Brewers. Brewer #016 is a single monochromator If a scan of the global irradiation from the dome is (MKII). His observational range is 290-325nm in treated, then the corresponding responsefile is applied steps of .5 nm. However before May 1991 the to convert the counts per seconds into Watt per software performed only scans in the range 290-315 square meter per nm. In this case the mean of the nm. A lamp unit and five 50W lamps were used to measured values at the 5 shortest wavelengths is keep track of the calibration level of the instrument. considered as a measure of the internal stray-light, The instrument participated in 1994 in the laboratory and is subtracted from the whole spectrum. This is campaign of CAMSUM at IASB [A. Webb, 1997]. In most important for the single monochromator. 1998 the calibration of the lamps was compared with offset and the correlation between the two data sets (not shown here). The differences at the lowest wavelengths are probably due to stray light problems in the single monochromator instrument. The UVI time series For the calculation of the UVI the measured spectrum is multiplied with the CIE action spectrum [CIE, 1987] and then integrated between 290 and 400nm. Since the instruments do not measure the entire UV-A region we have to estimate the contribution at wavelengths above the highest Figure 1. Relative deviation of the response at 5 wavelengths measured 315 or 325 nm for the single of Brewer #016 from the response curve in use at the monochromator and 363 for the double). This is done corresponding time. by completing the spectrum with an artificial spectrum that is matched with the mean value of the To account for changes in the response of the observations at the five longest observed instrument due to ageing or interventions, different wavelengths. Figure 3 shows the time series of the response files are used for different times. Between daily maximum values of the UVI’s obtained in this these times linear interpolation with time is way. The seasonal cycle arising from the change in performed between the calibration files. the elevation of the sun is obvious. Trends will be To get an idea of the remaining variation, Fig. 1 calculated for fixed solar zenith angles. shows the relative deviation of the response of Brewer 016 determined from individual lamp scans and the response spectrum in use at the corresponding time. It can be seen that the stability of the instrument obtained in this way better than +/- 5% (using 12 responsefiles over the 16 years). For the double mononochromater the results are even slightly better (+/- 3% with 2 calibration files over 3 years, not shown). Note that these comprise all different lamps that were used (seven different 50W as well as two 1000W lamps). Comparison of instruments A comparison of the spectra obtained with Brewer #016 (single monochromator) and Brewer #178 Figure 3. Daily maximum of the UVI at Uccle as measured (double monochromator) is shown in Fig. 2. It is seen with Brewer #016 (red asterisks) and Brewer #178 (green crosses). that both instruments compare very well above about 297nm. This was confirmed by the calculation of the Acknowledgements Part of this work was supported by the Belgian Federal Science Policy Office within the ESAC grant number II project EV/34/3B. References Bais, A. F., C. S. Zerefos, C. Meleti, I. C. Ziomas, and K. Tourpali, Spectral measurements of solar UVB radiation and its relation to total ozone, SO2, and clouds, J. Geophys. Res. (98), (D3) 5199-5204, 1993. CIE, A reference action spectrum for ultraviolet induced erythema in Human skin, CIE-Journal,6:17-22, 1987. Webb, A, Advances in solar ultraviolet spectroradiometry, Final report, European commission, Luxembourg, ISBN 92-828- 0990-0, 1997. WMO, The fifth biannual WMO consultation on Brewer Ozone and UV spectrophotometer operation, calibration and data reporting, edited by T. McElroy and E. Hare, WMO-GAW report nr 139, 1998. Figure 2. Mean ratio of the intensities measured within 1 minute of time with Brewer 178 and Brewer 016 between October 2001 and April 2004. The black line represents two standard deviations of the mean.
Pages to are hidden for
"595"Please download to view full document