J/A+A/705/A247 Custom ZTF photometry for classification (Arevalo+, 2026)
Unlocking AGN Variability with custom ZTF photometry for high-fidelity light
curves and robust selection.
Arevalo P., Sanchez-Saez P., Sotomayor B., Lira P., Bauer F.E., Rios S.
<Astron. Astrophys. 705, A247 (2026)>
=2026A&A...705A.247A 2026A&A...705A.247A (SIMBAD/NED BibCode)
ADC_Keywords: Active gal. nuclei ; Optical
Keywords: methods: data analysis - galaxies: active - galaxies: photometry -
quasars: general
Abstract:
We explored the potential of optical variability selection methods in
order to identify a broad range of active galactic nuclei (AGN),
including those challenging to detect with conventional techniques.
Using the unprecedented combination of depth, sky coverage, and
cadence of the Zwicky Transient Facility (ZTF) survey, we specifically
target low-luminosity, low-mass, and starlight-dominated AGN, known
for their redder colours, weaker variability signals, and difficult
nuclear photometry due to their resolved host galaxies.
We performed aperture photometry on ZTF reference-subtracted g-band
images for ∼39.8 million sources across >8000deg2; assembled light
curves and calculated features for all detected sources; and
classified objects employing a random forest algorithm into 14
distinct classes, including 341 938 candidate AGN across four classes
(low-z, mid-z, high-z, and blazars). We compared the variability
metrics derived from our photometry to those obtained from publicly
available ZTF data release light curves, obtained through
psf-photometry on the science, i.e. not reference-subtracted images
(DR11-psf), to assess the impact of our analysis. Finally, we compared
our AGN candidate sample with those identified through colour
selection and X-ray detection techniques.
We find that the fraction of low-z quiescent galaxies exhibiting
significant variability drops dramatically (from 98% of the sample to
7% of the sample, when using the standard variability metric Pvar)
when replacing the DR11-psf light curves with our difference image,
aperture photometry (DI-Ap) version. The overall number of variable
low-z AGN remains high (99% when using DR11-psf light curves, 83% when
using DI-Ap); however, it implies that our photometry can detect the
fainter variability in host dominated AGN. The classifier effectively
distinguishes between AGN and other sources, demonstrating high
recovery rates even for AGN in resolved nearby galaxies. AGN
candidates in eROSITA's eFEDS field, detected in X-rays and bright
enough for ZTF optical observations, were classified as AGN (79%) and
non-variable galaxies (20%). These groups show a 2dex difference in
X-ray luminosity, but not in X-ray flux. A significant fraction of
X-ray AGN are optically too faint for ZTF, and conversely, one-quarter
of ZTF AGN in the eFEDS area lack X-ray detections, highlighting a
wide range of X-ray-to-optical flux ratios in the AGN population.
Description:
Hierarchical random forest-based classification of astronomical
sources approximately within -30<dec<15 and gallat>20 and optical
g-band magnitude less than 20.5. The features used for the
classifications are variability features extracted from custom made
aperture photometry on Zwicky transient Facility difference images
from ZTF DR 13, in addition to Gaia proper motions, PanSTARRS and WISE
photometric data, and a morphology indicator. The first level
classifies sources as variable or non-variable, variable objects are
further classified into transient, stochastic or periodic, and each of
these is further classified into 15 classes of variable astronomical
sources. Non-variable objects are classified as either non-variable
star or non-variable galaxy.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
catalog.dat 164 39772280 Coordinates, some light curve features,
classifications and their probabilities
--------------------------------------------------------------------------------
Byte-by-byte Description of file: catalog.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 22 A22 --- OID Object identifier from this work
24- 38 F15.11 deg RAdeg Right ascension (J2000)
40- 54 F15.11 deg DEdeg Declination (J2000)
56- 60 F5.2 mag magMean Average magnitude
62- 64 I3 --- nEpochs Number of good epochs (catflags=0)
66- 72 F7.2 d TimeRange Total length of light curve used
74- 85 E12.6 --- Pvar Probability that the source is
intrinsically variable
87- 91 F5.3 mag Std Standard deviation of the light curve
93-104 E12.6 d GPDRWtau Relaxation time from DRW modeling
106-118 A13 --- predInitClass Predicted class from node_init
120-124 F5.3 --- predInitClassProb ?=- Predicted class probability from
node_init
126-135 A10 --- predVarClass Predicted class from node_variable
137-141 F5.3 --- predVarClassProb ?=- Predicted class probability from
node_init
143-156 A14 --- predClass Final predicted class
158-162 F5.3 --- predClassProb Final predicted class probability
164 I1 --- CleanAgnCand [0/1] Flag, 1: in clean AGN catalog,
0: not in clean AGN catalog
--------------------------------------------------------------------------------
Acknowledgements:
Patricia Arevalo, patricia.arevalo(at)uv.cl
(End) Patricia Vannier [CDS] 07-Nov-2025