Calibration of an instrument requires the determination of the calibration constant Fo and the effective ozone absorption coefficient A a as discussed in the previous section. Determination of A a is done for all instruments by applying laboratory measurements of the slit transmission functions to the ozone absorption coefficient spectrum. The absorption spectrum of Bass and Paur (1985) has been used operationally since this is used by other operational ground-based (Dobson) and satellite-based (Total Ozone Mapping Spectrometer (TOMS)) instruments. Other ozone absorption spectra are available (e.g., Molina and Molina, 1986; Brion et al., 1993), and if these spectra are applied to the Brewer instrument, differences of up to a few percent in total ozone result.
The slit transmission functions are determined by measuring the signal of emission lines at accurately known wavelengths as they are scanned individually across each exit slit by moving the micrometer. Emission lines from mercury, cadmium, neon, and other elements are used (Grobner et al., 1998). The known wavelengths are then used to determine the wavelength setting as a function of micrometer step for each slit, and the shape of the transmission as a function of wavelength is accurately determined for each slit. Operational measurements involving wavelength scans (e.g., global UV scans) use the wavelength dispersion function to select wavelength settings during the scans.
There are two methods to determine the extraterrestrial calibration constant (Fo from Eq. (6.3)): (1) the primary calibration method, and (2) the calibration transfer method. Primary calibrations are done by use of the Langley plot technique, also called the zero airmass extrapolation technique (Kerr et al., 1985). If the aerosol term (A^ (SZA)) in Eq. (6.3) is ignored then, the measured value (F), plus the calculated Rayleigh term (Afim) is a linear function of airmass with the value Fo as the intercept. Therefore, if measurements are made under a range of ^ (e.g., 1 < ^ < 3.5) throughout a day when ozone remains constant, then a plot of (F + AySm) versus ^ would yield a straight line with intercept Fo.
Extraterrestrial calibrations of Brewer instruments are carried out regularly at Mauna Loa Observatory (MLO), Hawaii (19.5°N, 155.6°W, 3400 m above sea level). This site offers stable observing conditions required for the calibrations and a low ^ value at noon (^ < 1.2) for most of the year. The high altitude of MLO is above most tropospherical contamination and the tropical location ensures minimal day-to-day variability of stratospheric ozone. In reality, ozone does not remain constant during the day at MLO and typically varies by about + 2 DU (1 - <r) throughout any given day (see Kerr, 2002), resulting in an uncertainty of Fo that would cause an error in total ozone of about 1.5%. Calibrations are therefore averaged over a 10-day period to reduce the effects of daytime ozone variability. The calibrations are done in two parts. The results of the first part are applied to past data and those of the second part are used for future data. Any changes or upgrades to the instrument (e.g., realignment, replacing or cleaning optical components, etc.) are carried out between the first and second parts.
The calibration reference for the global Brewer network is a triad of instruments based in Toronto. These instruments have been calibrated independently from the basic physical principles described above. There are three instruments to ensure the integrity and continuity of the reference. For example, if one instrument should develop a problem, the problem instrument is quickly identified as the outlier and measures are taken to correct the problem promptly. Historically since 1982, there has been at least one of the three instruments in operation at all times, fulfilling the objective of providing the continuous availability of a calibration reference. Normally, all three instruments are operating together and analysis of their long-term records shows agreement between the instruments comprising the triad better than 1% over a 20 year period (Fioletov et al., 2005).
The second method for determining Fo for an instrument is by the calibration transfer method, which uses an instrument with a known Fo. Simultaneous direct sun measurements are made with the calibrated instrument and one or more uncalibrated instruments. The total ozone value (O3) from the calibrated instrument is used to determine the Fo of the uncalibrated instrument using Fo = F + Afim + AaO3n (from Eq. (6.3) with the AS sec (SZA) set equal to zero). In practice, Fo is determined from the results of many side-by-side measurements made over the period of one or two days to provide a large range of airmass.
Calibration of operational field instruments is done by use of a traveling standard. This process uses an instrument, which is calibrated using the primary standard triad in Toronto. This traveling instrument is then transported to a site where it is used as the calibration reference for the field instrument. In practice, several regional field instruments are calibrated in one trip. After calibration of the field instrument(s), the traveling standard returns to Toronto for a follow-up calibration check. This operational process has the advantage that both the triad reference, as well as the field instruments, remain in operation continuously and are not subject to any risks involved with transporting an instrument.
Another method by which field instruments are calibrated is through intercomparison field campaigns. This method has been used traditionally for Dobson instruments, and Brewer instruments have often participated in these Dobson intercomparisons to ensure agreement between Brewer and Dobson references (e.g., Komhyr et al., 1989). Also, in some situations, it is more efficient for several field Brewer instruments to be calibrated by a traveling standard during a regional intercomparison campaign. Overall, the intercomparison campaign method is less desirable than the one-on-one site calibration method since the absence of field instruments during the intercomparison leaves gaps in the data records at field sites and instruments are subject to damage during transport to and from the intercomparisons.
Was this article helpful?