One subtlety is that the additive noise is unaffected by the atmospheric
decorrelation, in contrast with the signal, because thermal noise is a
random process as the turbulence phase noise. As the conversion factor,
, is applied to the data that can
contain signal as well as noise, any attempt to measure the noise rms on
visibilities or imaged data will thus result in a standard deviation
larger than the one given in Eq.
by a
factor
. So when we estimate the noise level of an interferometer,
we need to take into account the interferometric conversion factor that
depends on the typical weather conditions (i.e., the atmospheric rms
phase noise). This gives