A Hybrid Algorithm for Period Analysis from Multi-band Data with Sparse and Irregular Sampling for Arbitrary Light Curve Shapes

Ongoing and future surveys with repeat imaging in multiple bands are producing (or will produce) time-spaced measurements of brightness, resulting in the identification of large numbers of variable sources in the sky. A large fraction of these are periodic variables: compilations of these are of scientific interest for a variety of purposes. Unavoidably, the data-sets from many such surveys not only have sparse sampling, but also have embedded frequencies in the observing cadence that beat against the natural periodicities of any object under investigation. Such limitations can make period determination ambiguous and uncertain. For multi-band data sets with asynchronous measurements in multiple pass-bands, we want to maximally utilize the information on periodicity in a manner that is agnostic of differences in the light curve shapes across the different channels. Given large volumes of data, computational efficiency is also at a premium. This paper develops and presents a computationally economic method for determining periodicity which combines the results from two different classes of period determination algorithms. The underlying principles are illustrated through examples. The effectiveness of this approach for combining asynchronously sampled measurements in multiple observables that share an underlying fundamental frequency is also demonstrated.