The NN2 Flux Difference Method for Constructing Variable Object Light Curves

Abstract

We present a new method for optimally extracting point-source time variability information from a series of images. Differential photometry is generally best accomplished by subtracting two images separated in time, since this removes all constant objects in the field. By removing background sources such as the host galaxies of supernovae, such subtractions make possible the measurement of the proper flux of point-source objects superimposed on extended sources. In traditional difference photometry, a single image is designated as the ``template'' image and subtracted from all other observations. This procedure does not take all the available information into account and for sub-optimal template images may produce poor results. Given N total observations of an object, we show how to obtain an estimate of the vector of fluxes from the individual images using the antisymmetric matrix of flux differences formed from the N(N-1)/2 distinct possible subtractions and provide a prescription for estimating the associated uncertainties. We then demonstrate how this method improves results over the standard procedure of designating one image as a ``template'' and differencing against only that image.