Incidence of the host galaxy on the measured optical linear polarization of blazars

We study the incidence of the underlying host galaxy light on the measured optical linear polarization of blazars. Our methodology consists of the implementation of simulated observations obtained under different atmospheric conditions, which are characterized by the Gaussian sigma of the seeing function. The simulated host plus active nucleus systems span broad ranges in luminosity, structural properties, redshift and polarization; this allows us to test the response of the results against each of these parameters. Our simulations show that, as expected, the measured polarization is always lower than the intrinsic value, due to the contamination by non-polarized star light from the host. This effect is more significant when the host is brighter than the active nucleus, and/or a large photometric aperture is used. On the other hand, if seeing changes along the observing time under certain particular conditions, spurious microvariability could be obtained, especially when using a small photometric aperture. We thus give some recommendations in order to minimize both unwanted effects, as well as basic guidelines to estimate a lower limit of the true (nuclear) polarization. As an example, we apply the results of our simulations to real polarimetric observations, with high temporal resolution, of the blazar PKS 0521-365.