Aims. Analysing the long-term radio variability of active galactic nuclei (AGNs) is essential to understanding the physics of relativistic jets launched by supermassive black holes. We aim to connect the characteristic timescales obtained from a prior power spectral density (PSD) analysis to the decomposed timescales of the light curves. In addition, we probe for potential associations between the timescales and the physical characteristics of the relativistic jet as well as the central engine.

Methods. We decomposed the long-term radio light curves of 54 sources observed at the Aalto University Metsähovi Radio Observatory into individual flares to understand which timescale of variability is related to the low-frequency bend in the PSD. In addition, we used the obtained rise times of the brightest flares to look for associations between the emission-region size in the jet and different central engine parameters.

Results. We found that the inverse of the PSD bend frequency of radio light curves best corresponds to the mean duration of the brightest flares. For some sources, the mean flare separation had a similar timescale. Using the flare durations and separations as proxies for the PSD timescale, we found a positive correlation with black hole mass divided by the normalised mass accretion rate. This suggests that the variability timescales obtained from the PSDs of radio light curves are associated with the central engine. Furthermore, when comparing the obtained rise times of the brightest flares to the jet and central engine parameters, we found weak tentative correlations, but they may be driven by a common dependency on redshift.