Photometric light curve studies: potential bias induced by exposure time

TL;DR

This paper investigates how the common practice of averaging flux over exposure time in photometry can introduce biases in light curves, affecting the measurement of flux, peak positions, and colors, especially for short-timescale transients and variables.

Contribution

It reveals the potential biases caused by exposure time averaging in photometric light curves and discusses their impact on various astrophysical objects, proposing the need to account for this bias.

Findings

Exposure time averaging can flatten true light curves.

Peak flux and color measurements can be biased.

Bias affects flux estimates of stellar flares and AGN variability.

Abstract

In photometric observations, the flux averaged over the preset exposure time is usually used as the representation of an object's true flux at the middle of the exposure interval. For the study of transients and variables, it is also the default manner to build the light curves. In this work, we investigate the effect of this common practice on quantifying the photometric light curves. Our analysis shows that the flux averaged over the exposure time is not necessarily identical to the true flux so that potential bias may be introduced. The overall profile of the true light curve tends to be flattened by the exposure time. In addition, it is found that the peak position and photometric color can also be altered. We then discuss the impacts of the bias induced by exposure time on the light curves of stellar flares, periodic stars, and active galactic nuclei (AGNs). The bias can lead to an underestimate of the total fluxes of stellar flares which has been noticed in the observational data. For periodic stars that follow a sinusoidal light curve, the bias does not affect the period and peak position, but can result in the peak flux being underestimated. Meanwhile, the bias can result in steeper structure function at short timescales for AGN light curves. To obtain unbiased physical parameter estimates from the light curves, our analysis indicates that it is essential to account for this bias, particularly for transients and variables with very short timescales.