On the proper kinetic quadrupole CMB removal and the quadrupole anomalies

It has been pointed out recently that the quadrupole-octopole alignment in the CMB data is significantly affected by the so-called kinetic Doppler quadrupole (DQ), which is the temperature quadrupole induced by our proper motion. Assuming our velocity is the dominant contribution to the CMB dipole we have v/c = beta = (1.231 +/- 0.003) * 10^{-3}, which leads to a non-negligible DQ of order beta^2. Here we stress that one should properly take into account that CMB data are usually not presented in true thermodynamic temperature, which induces a frequency dependent boost correction. The DQ must therefore be multiplied by a frequency-averaged factor, which we explicitly compute for several CMB maps finding that it varies between 1.67 and 2.47. This is often neglected in the literature and turns out to cause a small but non-negligible difference in the significance levels of some quadrupole-related statistics. For instance the alignment significance in the SMICA 2013 map goes from 2.3sigma to 3.3sigma, with the frequency dependent DQ, instead of 2.9sigma ignoring the frequency dependence in the DQ. Moreover as a result of a proper DQ removal, the agreement across different map-making techniques is improved.