Photometric Biases in Modern Surveys

Precise brightness measurements, or photometry, are essential for many areas of astronomy. DiRAC Postdoctoral Fellow Stephen Portillo, together with collaborators Joshua Speagle and Douglas Finkbeiner, published the  paper “Photometric Biases in Modern Surveys” which illustrates a measurement bias that is common in astronomical surveys.

Surveys often use maximum likelihood methods to fit a model of the noise-free sky to observed images. Fitting the model gives measurements of the positions and brightnesses of astronomical objects like stars and galaxies. However, when the position of an object is not known beforehand but is part of the fit, Portillo and collaborators show that the brightness of that object will be systematically overestimated. This bias is strongest for the faintest objects, which are often the most interesting objects. Resolved objects like galaxies are more drastically affected by this bias. The authors also show that forced photometry, a common method to combine images taken with different telescope filters, can bias the measured color of sources.

Portillo and collaborators derive a formula that can be used to correct the bias and confirm that formula on simulated images. They then show that the bias exists in observations from the Sloan Digital Sky Survey and show that their formula predicts the bias consistently. Portillo & Speagle were recently interviewed about this paper by Frank Timmes, a Senior Lead Editor of the American Astronomical Society Journals: