Wavelet Analysis of OJ 287 As A Demo

Can continuous wavelet transforms reveal long-sought periodicities or other structure in the UMRAO database?

A Morlet transform of several decades worth of 14.5 GHz data for the source OJ 287:

What does this tell us? Is it white noise:

Not likely!!

Note that a wavelet transform provides clear evidence for periodicity, because of redundancy: the transform reflects the peaks and troughs of the signal, and shows this structure at a particular frequency (dilation); thus a pattern marches horizontally across transform space as seen in this period halving example:

OJ 287 has similar character, but it's not exactly like the previous example. How about the sum of two waves, at slightly different period, one modulated to dominate at mid-times:

Still not quite right! Here is a toy shock-in-jet model that does a strikingly good job at reproducing the behavior:

Compare with the first image above, noting that we are trying to reproduce the periodic structures, not longer and shorter time scale "noise".

This is a powerful technique for probing components of the signal. The periodicity is seen at all observing frequencies and in the polarized flux. That means we can "reconstruct" the Stokes parameters Q and U corresponding to the fluctuating signal and watch evolution in Q-U space with time:

MOVIE:

Time runs from the late 1970s (blue) to the present (red). The clear periodic excursions (mostly in U) suggest some sort of swinging or precession of the jet -- apparently unique to OJ 287, perhaps because in this source we see straight down the jet into its "throat" -- and constitute an exciting new way to explore these flows.