next up previous contents
Next: 3.2.3 Point Spread Function Up: 3.2 Galaxy Surface Photometry Previous: 3.2.1 Flat-Fielding   Contents


3.2.2 Sky Background

Even when the surface brightness contributed by the Sun and the Moon are not taken into account, the sky is never completely dark. One then has to subtract this contribution from the surface brightness measured in each region of the sky in order to derive the surface brightness contributed by the foreground object. In the optical region, the surface brightness of the moonless night sky, or Sky Background, results from several contributions, the most important of which are:

The total sky background, as well as the relative importance of its various components, vary greatly with wavelength, with time, with Galactic and ecliptic coordinates and from observatory to observatory. [Leinert et al. 1998] give a detailed account of the values of the sky background over a wide range of wavelengths, from the far ultraviolet to the far infrared. Generally, zodiacal light is the greatest contribution, followed by air glow, integrated starlight, diffuse Galactic light and diffuse extragalactic light. Table 3.1 gives typical values for the sky background as observed from the ground in the $ UBVRI$ bands.

Table 3.1: Typical values of the sky background as seen from the ground in the $ UBVRI$ bands. Surface brightness values expressed in mag/arcsec$ ^2$. From [Binney and Merrifield 1998].
$ \mu_U$ $ \mu_B$ $ \mu_V$ $ \mu_R$ $ \mu_I$
22.0 22.7 21.8 20.9 19.9

If seen from above the atmosphere, the sky background is fainter by a factor depending on the waveband. In the $ V$ band this factor can be as large as 1.5 mag, and still larger factors can apply at longer wavelengths. Specifically, [Gilmore 1997] has provided values of the sky background obtained excluding stars from some low Galactic and ecliptic latitude HST WFPC2 images. These values are listed in Table 3.2, and are believed to overestimate the typical sky background.

Table 3.2: Typical values of the sky background as seen from space in the $ BVI$ bands. Values are actually given in the HST bands F450W, F606W and F814W, closely resembling the $ BVI$ standard magnitudes (see [Holtzman et al. 1995b]). $ 1~\sigma$ random errors, derived from measurements at many different parts of the field, are of about 0.05 mag/arcsec$ ^2$. From [Gilmore 1997].
$ \mu_B~\mathrm{(F450W)}$ $ \mu_V~\mathrm{(F606W)}$ $ \mu_I~\mathrm{(F814W)}$
22.87 22.06 21.46

Note, however, that the sky background as measured by GAIA will be slightly brighter, due to unresolved objects, an effect which can conservatively be taken into account by increasing the values given in Table 3.2 by 0.5  mag. Note also that the real sky can be much brighter, due to scans through the Moon, bright planets, Galactic nebulae and very crowded regions. All such special cases need to be considered individually. In the following, specifically in Chapter 5, we will accordingly assume an $ I$-band sky background of $ \mu_{bg,I}=21.0$ mag/arcsec$ ^2$. Note that, in this context, the corrections to be applied for the conversion between HST WFPC2 magnitudes and standard $ UBVRI$ magnitudes, which in principle could be carried out using the HST WFPC2 calibration by [Holtzman et al. 1995b], are negligible.

The accuracy in the determination of the sky background is a critical factor in the final accuracy of galaxy surface photometry, and particularly for low-surface-brightness galaxies or in the outermost faint regions of otherwise high-surface-brightness galaxies. In these cases, one often has to follow the galaxy surface brightness distribution down to less than 1% of the sky background. Such measurements will clearly be meaningless unless the sky background can be extremely accurately determined, and small errors in the determination of the sky background can result in large errors of the derived galaxy surface brightness radial profile. Even more dramatic errors are to be expected in the radial distributions of galaxy color indices, since these have to be obtained by subtracting two derived surface brighness radial profiles.

Besides the already significant advantage of the lower sky background, space observations are relatively easy to correct for this effect, since they do not suffer from the rapidly varying (in space and time) air glow. On the other hand, the typically small field of view of space observatories does not generally allow a very accurate determination of the sky background in the surroundings of a diffuse object such as a galaxy. This is still a severe problem with CCD detectors, e.g. with HST observations. It is expected, however, that this will not be the case for GAIA, since the Astros, provided the necessary data can be accommodated into the telemetry, will be able to map the sky background over the whole sky at the desired resolution. It should also be emphasized that the dominant contribution to the measurement errors in GAIA surface photometry is the readnoise, with an essentially negligible contribution from the sky background.


next up previous contents
Next: 3.2.3 Point Spread Function Up: 3.2 Galaxy Surface Photometry Previous: 3.2.1 Flat-Fielding   Contents
Mattia Vaccari 2000-12-05