Automating the detection of polarization angle rotations in blazars
Anastasia Glykopoulou, Ioannis Liodakis, Dmitry Blinov · 2026
Read Automating the detection of polarization angle rotations in blazars by Anastasia Glykopoulou, Ioannis Liodakis, Dmitry Blinov (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-18955.txt.
Introduction
Anastasia Glykopoulou, Ioannis Liodakis and Dmitry Blinov automate finding polarization-angle rotations in blazars, replacing manual work. Their pipeline corrects the 180-degree ambiguity, segments light curves with Bayesian Blocks and checks each event. On RoboPol data it finds 48 rotations across 25 sources, including 27 new ones, with sizes from about 91 to 360 degrees and lengths from 7 to 111 days. Longer rotations tend to track stronger gamma-ray output, while size alone does not predict gamma-ray brightness.
Catalog record
| CREATOR | Anastasia Glykopoulou, Ioannis Liodakis, Dmitry Blinov |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36350 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.18955
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Automating the detection of polarization angle rotations in blazars Anastasia Glykopoulou, Ioannis Liodakis, Dmitry Blinov Abstract We present an automated pipeline for the detection of EVPA rotations in blazars, integrating correction of the 180 ∘ ambiguity, Bayesian Blocks segmentation, and statistical validation. Applied to RoboPol monitoring data, the method identified 48 rotations across 25 sources, including multiple events in RBPLJ2232+1143, RBPLJ1751+0939, RBPLJ1800+7828, and RBPLJ2253+1608. The rotations span amplitudes from 90.8 ∘ 90.8^{\circ} to 359.7 ∘ 359.7^{\circ} , durations between 7.0 and 111.3 days, and rotation rates averaging 5.0 ∘ 5.0^{\circ} /day. Comparison with previous catalogs reveals systematic differences: Bayesian Blocks rotations are on average ∼ 10 % \sim 10\% larger in amplitude, about twice as long in duration, and roughly two-thirds slower in angular velocity, reflecting systematic biases between adaptive binning and manual segmentation. In addition, we report 27 previously unreported rotations, including 11 from the final 2016–2017 season. A correlation analysis with contemporaneous Fermi–LAT γ \gamma -ray light curves shows that longer rotations tend to coincide with enhanced γ \gamma -ray activity, while rotation amplitude alone is not predictive of γ \gamma -ray brightness. Our pipeline minimizes subjective biases, expands the list of known EVPA rotations, and provides a reproducible framework for future multiwavelength studies of blazar jet dynamics and particle acceleration. Re-analysis of RoboPol data reveals 27 new rotations. 1 Introduction Blazars are a subclass of active galactic nuclei (AGN) whose powerful relativistic jets are oriented within a few degrees from the observer’s line of sight on Earth ( 5 ; 14 ) , producing bright and strong variability across the electromagnetic spectrum. The optical emission is dominated by synchrotron radiation from the jet, and therefore is highly polarized. Optical polarization monitoring programs have revealed that the electric vector position angle (EVPA) of the jet often undergoes large and systematic rotations, which provide a unique probe of the magnetic field structure and particle acceleration processes in jets. Such rotations have been associated with shocks ( 18 , e.g.,) , turbulence ( 19 , e.g.,) , or geometrical changes in the jet ( 1 ; 23 ; 12 ) , and have been found to statistically coincide with enhanced \gamma -ray activity ( 10 ) , suggesting a physical connection between polarization variability and high-energy emission. The first systematic study of EVPA rotations was carried out in the framework of the RoboPol program ( 3 ; 7 ; 9 ; 8 ; 10 ; 11 ) , which provided dense, multi-epoch polarimetric monitoring of a large blazar sample ( 22 ) . These efforts established the first statistical catalogs of EVPA rotations, employing well-defined algorithmic procedures for event identification. While effective, these methods remained sensitive to measurement uncertainty and intrinsic variability, which could influence the outcome. The highly stochastic nature of blazar jet variability ( 19 ; 15 ) therefore warrants more robust identification techniques. More recently, long-term monitoring of \gamma -ray blazars has emphasized the importance of reproducible methodologies to quantify such correlations across large samples ( 20 ; 24 ) . Motivated by 24 , we developed a fully automated pipeline tailored to the RoboPol blazar sample, capable of resolving the 180^{\circ} ambiguity, segmenting time series with Bayesian Blocks, and applying rigorous statistical tests. In this work we present the pipeline and its application to the RoboPol dataset, producing a reproducible catalog of rotation events, recovering previously overlooked cases, and extending the sample into the final RoboPol observing season 2016–2017 that had not been previously used to detect rotations. By comparing our results with earlier catalogs and examining correlations with contemporaneous \gamma -ray activity from Fermi–LAT ( 2 ) , we aim to provide new insights into the physical mechanisms driving EVPA rotations and their role in blazar jet dynamics. 2 Data and Methods 2.1 Data sample and observations The dataset used in this analysis originates from the RoboPol monitoring data described by 6 , which includes all sources of the main and control samples as well as individual sources observed with RoboPol over the duration of the program that were not part of the statistical analysis. Details on the monitoring program can be found in 22 & 3 , and the data analysis procedures in 16 ; 21 ; 6 . The monitoring campaign lasted from 2013–2017 and provides multi-epoch polarimetric measurements with dense temporal coverage, enabling the study of EVPA variability and long-term polarization behavior. In addition to the optical polarization data, we made use of the Fermi -LAT light curve repository ( 2 ) to examine the contemporaneous \gamma -ray activity of the sources and assess potential correlations with the optical polarization behavior. 2.2 Pipeline Structure 2.2.1 Preprocessing The monitoring data were ingested from CSV files containing the polarization measurements of each source. Julian dates were converted to Modified Julian Dates (MJD) for convenience. Time differences between consecutive observations were computed to facilitate the identification of observing gaps. To ensure statistical reliability, only measurements with a polarization degree significance of at least three ( \mathrm{PD}/\mathrm{err}_{\mathrm{PD}}\geq 3 ) were retained. The resulting dataset thus contains chronologically ordered, high-quality observations with consistent time stamps, forming the basis for subsequent segmentation and EVPA analysis. 2.2.2 EVPA 180^{\circ} Ambiguity EVPA measurements are defined modulo 180 ∘ , which can introduce artificial discontinuities in time series analysis. To avoid spurious rotations, we implemented an error-weighted adjustment procedure that explicitly accounts for EVPA measurement uncertainties e.g. ( 10 ) . Gaps larger than 30 days between consecutive observations prevent reliable angle disambiguation; such gaps are therefore used as natural segment boundaries (see also the section on rotation detection below). For each new data point, the measured angle was compared to the previously adjusted value, taking into account the combined uncertainty of the two measurements. If the difference was smaller than ninety degrees after weighting by the uncertainties, the angle was retained. Otherwise, integer multiples of 180 ∘ were added to the measured angle, and the candidate value that minimized the weighted difference with respect to the previous adjusted angle was selected. 2.2.3 Bayesian Blocks Analysis For the segmentation of the EVPA time series we employed the Bayesian Blocks algorithm ( 25 ) , using the implementation provided in the astropy.stats package ( 4 ) . The input consisted of the observation times (MJD), the adjusted EVPA values, and their measurement uncertainties. Since the EVPA data represent continuous measurements with Gaussian errors, we adopted the fitness=’measures’ option. The sensitivity to change points was controlled through the false-alarm probability parameter p_{0} , which we set to 0.001 . This configuration yields an adaptive segmentation of the EVPA time series into statistically homogeneous blocks, suitable for subsequent rotation and extrema analysis. 2.2.4 Identification of Local Extrema and Rotation Detection Algorithm The identification of EVPA rotation events was performed with a dedicated pipeline that integrates Bayesian Blocks segmentation, local-extrema analysis, and statistical validation. In contrast to the previous EVPA identification method used by RoboPol, our method does not impose smoothness constraints on the EVPA curve; candidate rotations are identified solely on the basis of statistically significant extrema pairs, regardless of the derivative behavior. As a result, events which would not have qualified as rotations under earlier criteria are included here. An example is shown