An insight into the extragalactic transient and variable microJy radio sky across multiple decades

TL;DR

This study investigates the faint, microJy extragalactic radio sky over two decades, revealing modest variability mostly linked to AGN activity, with short-term variability being more prevalent than long-term changes.

Contribution

First analysis of the microJy radio sky variability using deep VLA data across 22 years, highlighting the prevalence of short-term variability and its association with AGN.

Findings

Only up to 2% of sources show significant variability.

70% of variable sources vary on timescales of days.

80% of variable sources have VLBI counterparts.

Abstract

The mJy variable extragalactic radio sky is known to be broadly non-changing with approximately $3\%$ of persistent radio sources exhibiting variability which is largely AGN-related. In the faint (<mJy) flux density regime, it is widely accepted that the radio source population begins to change from AGN dominated to star-formation dominated, together with an emergent radio-quiet AGN component. Very little is known about the variable source component in this sub-mJy regime. In this paper, we provide the first insight into the $\mu$Jy variable sky by performing a careful analysis using deep VLA data in the well studied GOODS-N field. Using five epochs spread across 22 years, we investigate approximately 480 radio sources finding 10 which show signs of variability. We attribute this variability to the presence of an AGN in these systems. We confirm and extend the results of previous surveys, finding that variability in the faint radio sky is rather modest with only $\leq$2\% of sources exhibiting significant variability between any two epochs. We find that 70\% of variable sources show variability on timescales of a few days whilst on longer decadal time-scales, the fraction of variable sources decreases to $<1\%$. This suggests that the radio variability peaks on shorter timescales as suggested by other studies. We find that 80\% of variable sources have VLBI counterparts, and we use multi-wavelength data to infer that these may well be core-dominated FR-I sources as postulated by wide-field VLBI surveys and semi-empirical simulations.