9.7 Sky Background

As noted in the previous section, the backgrounds from the sky and from the detector must be taken into account when calculating expected signal-to-noise ratios (SNRs) or exposure times.

Detector backgrounds (as well as read-out noise) for the CCDs and IR detector are discussed in Chapter 5. Background added to CCD exposures using the post-flash option is discussed in Section 6.9.2. This section deals with the sky backgrounds that can affect WFC3 observations.

The most important sources of sky background are:

• Earthshine (ES).
• Zodiacal light (ZL).
  
• Geocoronal emission (GC).
  

The background in counts e⁻ s⁻¹ pixel⁻¹ for imaging observations can be computed as:

$$//<![CDATA[ \begin{array}{l}\displaystyle B_{sky} = \int{ I_{\lambda} S_{\lambda} m_{x} m_{y} d \lambda }\end{array} //]]>$$

where:

• I_λ is the surface brightness of the sky background, in erg cm⁻² s⁻¹ Å⁻¹ arcsec⁻².
• S_λ is the point-source sensitivity for the imaging mode.
  
• m_x and m_y are the plate scales in arcsec pixel⁻¹ along orthogonal axes.
  

In the case of slitless spectroscopy, the image of the sky through a disperser is not uniform, because some wavelengths fall off the detector for regions of sky near the edge of the field of view. The regions of lower sky background will be strips at the long- and short-wavelength edges of the field of view; a UVIS grism spectrum is roughly 270 pixels long, while an IR grism spectrum is roughly 170 pixels long. The maximum width of the strips from where the signal starts to decline to the edge, where the signal is down by roughly a factor of 2, will be about half the total length of a spectrum of a point source, i.e., roughly 135 pixels (UVIS) or 85 pixels (IR), in the case of a sky background with a continuum of wavelengths. These small strips of lower sky background are ignored in the following formulae. Furthermore, since the spectra do not lie along the direction of the anamorphic distortion, the plate scales of m_x and m_y  above must be replaced by the plate scales m_s and m_y  in the orthogonal spatial and dispersion directions, respectively. Interior to the strips, a point on the detector sees a region of sky over the full wavelength coverage of the disperser. Thus, for spectroscopic observations:

$$//<![CDATA[ \begin{array}{l}\displaystyle B^{\lambda}_{sky} = \int{ I_{\lambda} S'_{\lambda} m_{s} m_{\lambda} d\lambda }\end{array} //]]>$$

For a monochromatic sky emission line at λ = L like [O II] 2471, which will dominate the background through the UVIS/G280 grism:

$$//<![CDATA[ \begin{array}{l}\displaystyle B^{L}_{sky} = I_{L} S'_{\lambda} m_{s} \frac{m_{\lambda}}{d}\end{array} //]]>$$

where

• I_L is the monochromatic intensity of a line at wavelength L in erg cm⁻² s⁻¹ arcsec⁻².

The total sky background is:

$$//<![CDATA[ \begin{array}{l}\displaystyle B_{sky} = B^{\lambda}_{sky} + B^{L}_{sky}\end{array} //]]>$$

Figure 9.1 and Table 9.3 show "high" sky background intensity as a function of wavelength, identifying the separate components which contribute to the background. The "shadow" and "average" values of the earth-shine contribution in the WFC3 Exposure Time Calculator (ETC) correspond, respectively, to 0% and 50% of the "high" values in Figure 9.1 and Table 9.3. For the zodiacal sky background, the values in Figure 9.1 and Table 9.3 correspond to the high value of V-band surface brightness of 22.1 mag arcsec⁻² from Table 9.4, while the "low" and "average" zodiacal light is scaled to V surface brightnesses of 23.3 and 22.7 mag arcsec⁻², respectively.

In Table 9.3 we present the "high" sky-background numbers, which are plotted in Figure 9.1. See the text and the caption of Figure 9.1 for more details. These high sky values are defined as the earth-shine at 38° from the limb and the high zodiacal light of V = 22.1 mag arcsec⁻².

Table 9.3: High Sky Backgrounds.

9.7.1 Zodiacal Light, Earth Shine, and LOW-SKY

Zodiacal light varies as a function of angular distance of the target from the Sun and from the ecliptic. Table 9.4 gives the variation of the zodiacal background as a function of heliocentric ecliptic longitude and ecliptic latitude in V-mag per arcsec². The entries in the table are derived from data in Leinert et al. (1998). As shown by this table, over most of the sky the zodiacal background is within 1 magnitude of the minimum value of 23.4 V-mag per arcsec². It is greater than that at low ecliptic latitudes and at angles from the sun approaching the minimum permitted observing angle of 55°. At the lower ecliptic latitudes, it rises rapidly as this 55° limit is approached. For a target near heliocentric ecliptic coordinates (50,0) or (−50,0), not shown in the table, the zodiacal light level is 20.9 V-mag per arcsec², i.e., about 10 times the faintest value. For an empirical determination of the dependence of zodiacal light on the ecliptic latitude and the angle between the target and the sun for WFC3/IR observations, see Section 7.9.5. Figure 7.12 illustrates this dependence.

Table 9.4: Approximate zodiacal sky background as a function of heliocentric ecliptic longitude and ecliptic latitude (in V-mag per arcsec²). SA stands for Solar Avoidance zone (HST pointing within 55 deg of the sun), where observations may not be made.

Earth-shine varies strongly depending on the angle between the target and the bright Earth limb. The variation of the earth-shine as a function of limb angle from the sunlit Earth is shown in Figure 9.2. The figure also shows the contribution of the moon, which is typically much smaller than the zodiacal contribution, for which the upper and lower limits are shown. For reference, the limb angle is approximately 24° when the HST is aligned toward its orbit pole (i.e., the center of the CVZ). The earth-shine contribution shown in Figure 9.1 and Table 9.3 corresponds to this position.

For observations taken longward of 3500 Å, earth-shine dominates the background at small (<22°) limb angles. In fact, the background increases exponentially for limb angles <22°. The background near the bright limb can also vary by a factor of ~2 on timescales as short as two minutes, which suggests that the background from earth-shine also depends upon the reflectivity of the terrain and the amount of cloud cover over which HST passes during the course of an exposure.

Observations of the faintest objects may benefit from the use of the Phase II Special Requirement LOW-SKY, which must be requested and justified in the Phase I proposal. LOW-SKY observations are scheduled during the part of the year when the zodiacal background light is no more than 30% greater than the minimum possible zodiacal light for the given target position.  LOW-SKY also invokes the restriction that exposures will be taken only at angles greater than 40° from the bright Earth limb to minimize earth-shine and the airglow lines (but also see Section 9.7.2). However it is not always effective in suppressing the He I 10830 Å airglow line (Section 7.9.5).

One disadvantage of using the  LOW-SKY requirement is that it greatly reduces scheduling opportunities due to the restriction on zodiacal light. That restriction may be stricter than scientifically necessary, especially for targets at high ecliptic latitudes where even the maximum attainable zodiacal background is quite low, and for filters where other background components dominate the noise.

For targets and filters where high zodiacal background could be a problem, observers may want to limit that background by placing their own less stringent restrictions on scheduling dates to avoid observing the target when it is too near the sun. HST is already constrained to point at least 55° from the sun. For targets at low ecliptic latitude, that limit can be extended to ~60° by excluding 5 days from the ends of the default solar exclusion window, as described in Section 7.9.5.

Another disadvantage of using the  LOW-SKY requirement is that it limits visibility to about 48 minutes per orbit. Rather than accept an automatic reduction in observing time by using  LOW-SKY, the observer may be able to structure an orbit so that the exposures that would be most significantly affected by earth-shine do not occur at the beginning or end of the orbit, where the earth-shine is likely to be highest. The exposures in that orbit should be placed in a non-interruptible exposure sequence, as described in the APT Training Materials. For WFC3/IR exposures, frames in the non-destructive readouts that are badly affected by earth-shine can be excluded from the data reduction (see Section 7.9.5).

WFC3/UVIS observers should keep in mind that the total background (dark + sky + post-flash if needed) must be at least 20 electrons/pix/s to minimize CTE losses, so for some targets and filters it may be advisable to increase the background using post-flash rather than to attempt to minimize it (Section 6.9).

The ETC provides the user with the flexibility to separately adjust the zodiacal and earth-shine sky background component levels to help observers decide where special care is needed to avoid background levels that could be detrimental to their science. The RA, Dec and date option can be used to compute the zodiacal light for specific targets, which is especially useful for targets at low ecliptic latitudes where the zodiacal background increases rapidly as the angle between the target and the sun approaches the HST observable limit of 55°. Note that the standard normalizations for zodiacal light and earth-shine (from none to high or extremely high) provide approximations that may not accurately predict the levels in individual exposures. For WFC3/IR exposures, one can check the empirically determined zodiacal light model discussed in Section 7.9.5.

9.7.2 Geocoronal Emission, Airglow, and SHADOW

Background due to geocoronal emission originates mainly from hydrogen and oxygen atoms in the exosphere of the Earth. In the far-UV spectral region, the strongest geocoronal emission lines are Lyman−α at 1216 Å, O I at 1304 Å, and O I] at 1356 Å, but WFC3 is of course not sensitive at these wavelengths. The only significant UV geocoronal emission line to which WFC3 is sensitive is [O II] 2471 Å, shown in Figure 9.1. This line varies in intensity from day to night and as a function of HST orbital position. It can be the dominant sky contribution in the UV both for imaging and spectroscopic observations. In sunlight it can be as bright as ~1.5 × 10⁻¹⁵ erg cm⁻² s⁻¹ arcsec⁻², while in Earth shadow it is much fainter, typically ~7.5 × 10⁻¹⁸ erg cm⁻² s⁻¹ arcsec⁻². However, note that WFC3 ISR 2023-06 did not find any contribution attributable to [O II] to the G280 grism sky background.

To minimize geocoronal emission, the Special Requirement SHADOW can be used if requested and justified in the Phase I proposal. Exposures using this special requirement are limited to roughly 25 minutes per orbit, exclusive of the guide-star acquisition (or re-acquisition), and can be scheduled only during a small percentage of the year. SHADOW therefore restricts scheduling opportunities and orbit duration much more severely than LOW-SKY (see Section 7.9.1). SHADOW reduces the contribution from the geocoronal emission lines by roughly a factor of ten while the continuum earth-shine is set to zero. SHADOW requirements must be included and justified in your Phase I proposal (see the Call for Proposals). Observers should keep in mind that the total background should be at least 20 electrons/pix/s to minimize CTE losses in WFC3/UVIS images, so it may be advisable to increase the background using post-flash rather than to attempt to minimize it. (See Section 6.9)

Analysis of archival WFC3/IR grism and broad-band filter observations has shown that a variable airglow line of He I at 10830 Å can be a significant component of the sky background for the F105W and F110W filters and the G102 and G141 grisms. See Section 7.9.5 for discussion of the WFC3/IR sky background and observing strategies. The use of SHADOW is generally not recommended for WFC3/IR exposures.