2013MNRAS.431.1048K

A common problem in CCD photometry is to combine measurements obtained at different epochs. This is especially challenging if the fields only overlap partially, and are sparse, so that not all calibrating stars are observed at all epochs. A least-squares method for determining zero-points for all epochs under these circumstances is formulated. Allowance is made not only for the presence of intra-night measurement errors, but also for epoch-to-epoch scatter in star brightnesses (due, for example, to slow variability or instrumental effects). Expressions are derived for the uncertainties in the estimated zero-points. Three different criteria for selecting optimal calibrating subsets of the available group of stars are introduced. Simulations show that if intra-night variability dominates, it may be best to use all available stars for zero-point determinations. On the other hand, if epoch-to-epoch scatter dominates, smaller subsets of stars may give superior results. Characterization of the uncertainty in the estimated variance of epoch-to-epoch variability remains an important outstanding problem.