Line vs continuum system temperature

Figure [*] shows the system temperature as a function of the radio frequency. The $\ensuremath{T_\ensuremath{\mathrm{sys}}}$ can vary significantly over the large bandwitdh of the 2SB NOEMA receivers. Figure [*] shows the minimum and maximum system temperature inside the IF bandwidth for all possible local oscillator tunings. As a result, for continuum estimation, a frequency averaged $\ensuremath{T_\ensuremath{\mathrm{sys}}}$ is interpolated from a pre-computed table. The relevant frequency in that case is the LO frequency of the tuning (see Fig. [*]). The averaging is done such as

$\displaystyle \boxed{%
\frac{1}{<\ensuremath{T_\ensuremath{\mathrm{sys}}}>^2} = \frac{1}{N} \sum \frac{1}{\ensuremath{T_\ensuremath{\mathrm{sys}}}^2}.
}$ (19)