Subsections

• SSO CCD System
• Extinction Coefficients

Atmospheric Extinction

SSO CCD System

All the observational materials were obtained at SSO with the 40 inch telescope (f/8) and a thinned SITe 2048 x 2048 CCD (24 pixels). Across the U passband, the SITe CCD shows a rapid change in sensitivity (QE 60% at

and 10% at

). The specifications of the filters used in the observations is summarized in Table 1.

The response function of the U filter used in the CCD photometry is similar to the combined response of the U filter and 1P21 photomultiplier tube of the original Johnson U but the steeply sloping QE variation across the U passband results in an overall CCD U band response function significantly different from the standard U passband. Sung et al. (1998) included a non-linear correction term (f[(B-V)₀]) in the U transformation to the Landolt standards to account for the Balmer Jump as well as a linear term in U-B to correct for the effective wavelength shift. The transformation equation they used is,

where U, (U-B) and (B-V) denote standard values, u, (u-b) and (b-v) are intrumental values, k_1U is the primary extinction coefficient for U, X is the airmass and U, is the zeropoint in U. From 1997 May, for the U observations we added an additional 1mm UG1 filter to the existing U filter in order to shift the effective wavelength toward the UV and closer to that of standard U. In addition, to quantify what part of the non-linear correction was due to the Landolt U-B system (see Menzies et al. 1991; Cousins 1984a,b; see also Bessell 1995), we observed the SAAO E-regions. The number of usable data for the determination of the atmospheric extinction coefficients are given in Table 2; the coefficients are given in Fig. 1.

Extinction Coefficients

Figure 1: Measured Siding Spring extinction coefficients plotted against yearly day number. The weighted mean values are indicated by the line and the value listed in the figure. The square and circle represents, respectively, the data obtained in 1996 and 1997. The open and filled symbols in k_1U and k_2U denote without and with the additional 1mm UG1 filter, respectively.

Atmospheric extinction is caused by the scattering or absorption of light by molecules or other particles. Most of the extinction in the visual window is due to Rayleigh scattering by air molecules. Another important contributor to the extinction is scattering and absorption by small liquid or solid particles of various sizes called aerosols (Cousins & Caldwell 1998). In the case of an observatory in an urban area or at a low altitude non-urban site, extinction by aerosols is the most important source although the constituent aerosols will differ. Volcanic eruptions and massive forest-fires for example, can inject aerosols into the stratosphere resulting in large, wide spread and long lasting increases in extinction. Changes in the thickness of the ozone layer in the upper atmosphere also affects the extinction in the UV.

As the atmospheric extinction coefficients change mainly with changing aerosol content, in the absence of volcanic eruptions, dust storms or fires etc., the extinction coefficients are generally quite stable and the use of mean extinction coefficients is recommended so long as the standard stars and program objects are measured interspersed and at similar zenith distances (ZDs). That is, if standards and targets have similar ZD values (i.e. differ by less than 0.1), deviations from mean extinction (say by 0.1 mag, an unusually large difference) produce only a second order effect (less than 0.01 mag) on the accuracy of the photometry. But if the difference in airmass between standard stars and targets is large (say 0.5), then a 0.1 mag difference in the true extinction from the mean extinction will result in a difference of 0.05 mag in the photometry, an unacceptable error.

We determined atmospheric extinction coefficients under photometric conditions. By observing standard stars from the meridian to

(i.e. airmass ) we obtain a long baseline in airmass and thus measured precise extinction. In V, R, and I, there was no evidence for secondary extinction coefficients (k₂s) so only the primary extinction coefficients (k₁s) were derived. But in U and B, secondary extinction coefficients were evident. We determined and used weighted mean value of secondary coefficients for the atmospheric extinction corrections.

We plot the derived extinction coefficients obtained in 1996-97 in Fig. 1. Usually we observed only UBVI and therefore only a few k_1R values are shown. The large scatter in the secondary extinction coefficients was caused mainly by a lack of blue stars in the observed regions. The weighted mean values (weight =

, where is the standard error of coefficient) are marked in the figure.

The fluctuations in k₁ are up to 0.05 for a given filter. The use of an additional UG1 1mm filter in U observation did not appear to cause any noticeable difference in the U extinction. The mean values of the primary extinction coefficients are very similar to those obtained by Landolt at Cerro Tololo (see Table 1 of Landolt 1992). Only I showed a seasonal variation in extinction, being lower in winter. On 1997 August 9, the first third of the night showed normal extinction; it was non-photometric in the second part and the last part yielded a small value of k_1I. Probably variations in the column density of H₂O (strong absorption bands of H₂O fall within the I band) caused the variations in the I extinction coefficient.

© Copyright Astronomical Society of Australia 1997