Searching for Short-Timescale Variability in the Ultraviolet with the GALEX gPhoton Archive I.: Artifacts and Spurious Periodicities

TL;DR

This study investigates short-timescale ultraviolet variability in GALEX data, identifying and characterizing instrumental artifacts that can mimic real astrophysical signals, thereby improving the reliability of UV variability searches.

Contribution

The paper introduces a methodology to detect and analyze instrumentally-induced artifacts in GALEX UV photometry, highlighting previously unreported variability patterns and proposing correction strategies.

Findings

Identified quasi-sinusoidal artifacts mimicking pulsators with amplitudes up to 0.3 mag.

Discovered artifacts related to dithering patterns causing flux variations.

Provided methods to distinguish and correct instrumental artifacts in UV light curves.

Abstract

In order to develop and test a methodology to search for UV variability over the entire GALEX database down to the shortest time scales, we analyzed time-domain photometry of $\sim5000$ light curves of $\sim300$ bright $(m_{\rm{FUV}}, m_{\rm{NUV}} \leq 14)$ and blue $(m_{\rm{FUV}} - m_{\rm{NUV}} < 0)$ GALEX sources. Using the \gphoton \ database tool, we discovered and characterized instrumentally-induced variabilities in time-resolved GALEX photometry, which may severely impact automated searches for short-period variations. The most notable artifact is a quasi-sinusoidal variation mimicking light curves typical of pulsators, seen occasionally in either one or both detectors, with amplitudes of up to 0.3 mag and periods corresponding to the periodicity of the spiral dithering pattern used during the observation (P$\sim$120 sec). Therefore, the artifact may arise from small-scale response variations. Other artifacts include visit-long "sagging" or "hump" in flux, occurring when the dithering pattern is not a spiral, or a one-time change in flux level during the exposure. These instrumentally-caused variations were not reported before, and are not due to known (and flagged) artifacts such as hot spots, which can be easily eliminated. To characterize the frequency and causality of such artifacts, we apply Fourier transform analysis to both light curves and dithering patterns, and examine whether artificial brightness variations correlate with visit or instrumental parameters. Artifacts do not correlate with source position on the detector. We suggest methods to identify artifact variations and to correct them when possible.