12.4 Scintillation Noise
175
automatically plate solve a defocused image, as the plate-solving algorithm will not
be able to identify the stars in the field, as they are no longer Gaussian. Likewise,
any automatic photometry is also going to be impossible for the same reason. It is
also likely that you will not be able to defocus during remote or robotic operations.
Defocused photometry is more susceptible to bad and poor pixels (as the region
of the CCD/CMOS is larger) and larger dark noise due to the extended exposure
time, although this may be offset by lower scintillation noise. It also is not suitable
for remote and robotic observations, as they require the telescope to remain in focus.
12.5 Sky Noise
Sky noise is caused by changes in the emissivity of the atmosphere. Although the
atmosphere is largely transparent in the optical range it still emits a small amount
of light as well as reflecting light from the ground and astronomical objects from
dust particles in the atmosphere. In a well calibrated image, devoid of bad pixels,
with zero contribution from bias, dark, and flat-field sources, the count away from a
source represents the sky signal. You should notice when observing when the Moon
is up or when imaging close to a bright planet that your sky signal is higher.
As explained in Chap. 10, when performing aperture photometry, we sample the
sky signal using an annulus around the target (or away from the target if there is
another object in the annulus) and use the mean value of the sky in the annulus as
the sky value, which in turn is subtracted from the source count. The variation in the
sky signal is the sky noise. Sky noise is the main limiting factor in observing faint
objects. Sky noise is broadly equal to the root sky count, as sky noise is generally
Gaussian, especially for longer exposures. Hence, increasing the exposure time does
not improve the impact of sky noise on the uncertainties in your photometry. If your
target signal does not exceed the sky noise by at least a factor of two, you are unlikely
to achieve reliable detection.
When you perform aperture photometry using an application such as APT or
Maxim DL, the reported uncertainty produced by the software is based on the sky
count. The software assumes that noise from other sources has been addressed,
although in fact, in most cases it has not. Effectively, the software is giving you a
“best case” uncertainty. There are a few simple methods that can be undertaken to
reduce sky noise. The most obvious one is to site the observatory somewhere dark
and away from local population centres. This is not that practical for many teaching
observatories where the campus is in a big city. Almost all London-based universities
that teach astrophysics have sky noise issues from city lights as the city has grown
around them. Sky noise also scales with altitude, which is why many of the world’s
best observatories are at high locations, and the Hubble Space Telescope can perform
so well despite having only a 2.4 m mirror.
Observing your target at low air mass and when there is no Moon or when the Moon
is at a crescent phase will also improve sky noise, as will avoiding bright stars and
planets. High cirrus clouds, invisible from the ground, also increase sky brightness.
Précédent

- 183/242

Suivant