In the interferometric imaging, the visibility data averaged over a certain frequency range are treated as if they were measured with a monochromatic receiving system tuned to the center frequency of the range. This results in a radial smearing of the intensity distribution. This smearing effect becomes more severe with increasing fractional averaging bandwidth (frequency averaging width / observing frequency) and increasing the ratio of the offset from the primary beam center to the synthesized beam width (see Interferometry and Synthesis in Radio Astronomy by A. R. Thompson et al. for more details). The flux reduction factor (Rb) due to the smearing is given by the following equation.
Here θb is the synthesized beam width, ν0 is center observing frequency, r1 is the offset from the beam center, and Δν is the averaging bandwidth. The flux reduction becomes nonnegligible in Band 1 with long baseline configurations (C-8, 9, and 10). Users are advised to pay particular attention to continuum imaging in TDM. Fig 1 and 2 show the flux reduction factor with 31.25 MHz bandwidth, which is equivalent to an effective channel width of TDM, at 35 GHz and 50 GHz. The worst case is that flux density of the source near the edge of the primary beam is reduced by ~10% at 35 GHz (see black line in Fig 1). The flux reduction is less severe for FDM of 1.875 GHz bandwidth, but it may not be negligible when averaging over >~16 MHz (Fig 3).
For the observations that require a mapping over a wide area within the primary beam in Band 1, please consider using FDM of 1.875 GHz bandwidth with a spectral averaging of 4 (Fig. 4) or fewer channels. Note that the smearing effect is even worse for full polarization case since the width of channel becomes larger by a factor of two. For this case, please use FDM with a spectral averaging of 2 channels or without averaging.
Fig.1 The flux reduction factor (vertical axis) as a function of the distance from the beam center normalized by the primary beam radius (horizontal axis) at 35 GHz. The channel width is assumed to 31.25 MHz (effective channel width in TDM with Hanning smoothing).
Fig. 2 The same as Fig 1 but at 50 GHz.
Fig. 3 The same as Fig 1 but the channel width is assumed to be 16 MHz (close to the effective frequency resolution of FDM 1.875 GHz bandwidth with 8-channel averaging).
.
Fig. 4 The same as Fig 1 but the channel width is assumed to be 7.808 MHz (the effective frequency resolution of FDM 1.875 GHz bandwidth with 4-channel averaging).