Long-term NIR Variability in the UKIDSS Ultra Deep Survey: a new probe of AGN activity at high redshift

We present the first attempt to select AGN using long-term NIR variability. By analysing the K-band light curves of all the galaxies in the UKIDSS Ultra Deep Survey, the deepest NIR survey over ~1 sq degree, we have isolated 393 variable AGN candidates. A comparison to other selection techniques shows that only half of the variable sources are also selected using either deep Chandra X-ray imaging or IRAC colour selection, suggesting that using NIR variability can locate AGN that are missed by more standard selection techniques. In particular, we find that long-term NIR variability identifies AGN at low luminosities and in host galaxies with low stellar masses, many of which appear relatively X-ray quiet.