2010MNRAS.405..718P

We assess the effects of simulated active galactic nuclei (AGN) on the colour and morphology measurements of their host galaxies. To test the morphology measurements, we select a sample of galaxies not known to host AGN and add a series of point sources scaled to represent specified fractions of the observed V-band light detected from the resulting systems; we then compare morphology measurements of the simulated systems to measurements of the original galaxies. AGN contributions >rsim 20 per cent bias most of the morphology measurements tested, though the extent of the apparent bias depends on the morphological characteristics of the original galaxies. We test colour measurements by adding to non-AGN galaxy spectra a quasar spectrum scaled to contribute specified fractions of the rest-frame B-band light detected from the resulting systems. A quasar fraction of 5 per cent can move the NUV - r colour of an elliptical galaxy from the ultraviolet-optical red sequence to the green valley, and 20 per cent can move it into the blue cloud. Combining the colour and morphology results, we find that a galaxy/AGN system with an AGN contribution >rsim 20 per cent may appear bluer and more bulge dominated than the underlying galaxy. We conclude that (1) bulge-dominated, E/S0/Sa, and early-type morphology classifications are accurate for red AGN host galaxies and may be accurate for blue host galaxies, unless the AGN manifests itself as a well-defined point source and (2) although highly unobscured AGN, such as the quasar used for our experiments, can significantly bias the measured colours of AGN host galaxies, it is possible to identify such systems by examining optical images of the hosts for the presence of a point source and/or measuring the level of nuclear obscuration.