Fast Bayesian spectral analysis using Convolutional Neural Networks: Applications over GONG H α \alpha solar data

G. Castelló, M. Luna, J. Terradas · 2025

Read Fast Bayesian spectral analysis using Convolutional Neural Networks: Applications over GONG H α \alpha solar data by G. Castelló, M. Luna, J. Terradas (2025) 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-2501-12743.txt.

Introduction

Castelló, Luna, and Terradas teach neural nets to spot sun filament swings in a 2025 paper. Filaments are cool gas arcs on the sun that sway with set beats. Daily watch of those beats may tell how the arcs grow. Hand reads held the field back. The team first fit the sway spectra with a Bayesian chain method, the gold path. That path is slow, so they trained two nets to match it.

Each pixel of a sun watch cube gets a spectrum read. Noise in the read splits to red and white parts, both fit in the chain. The nets then call real swings at a set trust line. On past known cases the nets found the same events as the gold path. A fresh day with new events tried their reach. Run time fell by a wide span with no loss of trust.

Catalog record

CREATORG. Castelló, M. Luna, J. Terradas
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-12743

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Fast Bayesian spectral analysis using Convolutional Neural Networks: Applications over GONG H α \alpha solar data G. Castelló, M. Luna, J. Terradas Abstract Context. Solar filament oscillations have been observed for many years, but recent advances in telescope capabilities now enable daily monitoring of these periodic motions, offering valuable insights into the structure of filaments. A systematic study of filament oscillations over the solar cycle can shed light on the evolution of the prominences. Until now, only manual techniques have been used to analyze these oscillations. Aims. This work serves as a proof of concept, aiming to demonstrate the effectiveness of Convolutional Neural Networks (CNNs). These networks automatically detect filament oscillations by applying power-spectrum analysis to H α \alpha data from the GONG telescope network. Methods. The proposed technique studies periodic fluctuations of every pixel of the H α \alpha data cubes. Using the Lomb Scargle periodogram, we computed the power spectral density (PSD) of the dataset. The background noise is well fitted to a combination of red and white noise. Using Bayesian statistics and Markov chains - Monte Carlo (MCMC) algorithms we can fit the spectra and determine the confidence threshold of a given percentage to search for real oscillations. We built two CNN models to obtain the same results as the MCMC approach. Results. We applied the CNN models to some observations reported in the literature, proving its reliability in detecting the same events as the classical methods. A day with events not previously reported has been studied to check for the model capabilities outside of a controlled dataset where we can check with previous reports. Conclusions. CNNs proved to be a useful tool to study solar filament oscillations, using spectral techniques. Computing times have been reduced significantly while getting results similar enough to the classical methods. This is a relevant step towards the automatic detection of filament oscillations. 1 Introduction Solar filaments—also known as prominences when observed off-disk—are dense, cool plasma structures suspended in the Sun’s hot corona by magnetic fields, appearing as dark structures in H \alpha observations. Despite extensive research, the formation, structure, and disappearance of these filaments are not fully understood. They form within filament channels (FCs), which are large, non-potential magnetic structures situated above the polarity inversion lines (PILs) in the photosphere. The development of FCs is influenced by solar surface and subsurface motions, including differential rotation, meridional flows, and convection ( Mackay 2014 , see, e.g.,) . Throughout the solar cycle, filaments migrate across different latitudes, forming patterns that resemble the sunspot butterfly diagram ( McIntosh 1972 ; Minarovjech et al. 1998 ; Coffey & Hanchett 1998 ; Diercke et al. 2024 ) . This migration suggests a relationship between filament structures and the large-scale solar magnetic field. Solar prominences exhibit various types of motion, notably oscillations, which have become more frequently observed due to the capabilities of new instruments ( Arregui et al. 2018 ) . These oscillations are relevant because they enable prominence seismology to determine physical parameters of magnetized plasma structures. Oscillations in solar filaments are categorized based on their restoring forces. Transverse oscillations are mainly governed by the Lorentz force ( Hyder 1966 ; Kleczek & Kuperus 1969 ; Hershaw et al. 2011 ) , while longitudinal oscillations are influenced by gravity along magnetic field lines and gas pressure gradients ( Luna & Karpen 2012 ; Luna et al. 2012 ; Zhang et al. 2012 ; Zhang et al. 2013 ) . Luna et al. 2017 compared field properties derived from filament seismology with the reconstructed magnetic field using the flux rope insertion method by van Ballegooijen 2004 and found a general agreement. This highlights the potential of oscillation studies to enhance our understanding of filament structures. Despite the potential insights that systematic studies of filament oscillations could offer into the evolution of prominences and their FCs over the solar cycle, most observational research has focused on individual events. One of the few extensive studies was conducted by Bashkirtsev & Mashnich 1993 , who reported 38 oscillation events over nearly a solar cycle and noted a sinusoidal latitudinal dependence of oscillation periods. However, a later study by Luna et al. 2018 did not find a clear relationship between oscillation periods and filament latitude. Using H \alpha data from the GONG network during January–June 2014, Luna et al. 2018 provided an extensive sample of 196 events near the solar maximum of cycle 24 and found an average oscillation period of 58 minutes, regardless of amplitude or filament type. Traditionally, the time-distance technique has been used, which involves measuring the intensity along a fixed artificial slit following the motion. This method has been widely applied in the study of oscillations in coronal loops and the detection of solar filament oscillations ( Pant et al. 2016 ; Zhang et al. 2017 ; Joshi et al. 2023 , see) . However, these methods require firstly visual detection of the oscillations and secondly tracing an artificial slit to construct the time-distance diagram. This makes time-distance diagram methods not applicable for the automatic detection and parameterization of oscillations in filaments. Luna et al. 2022 introduced a new method by using two-dimensional H \alpha images and performing spectral (frequency-domain) analysis on the intensity time series of all pixels in the data. This technique allows the automatic detection of oscillations in the solar disk. Luna et al. 2022 demonstrated that spectral techniques can effectively detect oscillations in solar filaments, serving as a proof of concept. However, the methods they employed offer room for refinement. Following the approach of Auchère et al. 2016 , background noise is typically estimated by fitting the PSD using a weighted least squares method. This fit is weighted by the inverse square of the smoothed PSD, obtained via a five-point boxcar average. Additionally, confidence levels are calculated assuming that the background noise is white noise. The PSD derived from the temporal evolution of pixel intensities is usually a combination of red and white noise. Analyzing such PSDs is challenging because red noise can mimic or obscure genuine oscillatory signals, potentially leading to false detections if not properly accounted for. Sophisticated statistical methods are therefore necessary to distinguish true oscillations from noise-induced fluctuations. While some studies have attempted to simplify the analysis by whitening the PSD to mitigate the red noise component, VanderPlas 2018 emphasizes that a more rigorous approach involves following the Bayesian methodology outlined by Vaughan 2010 . This methodology employs MCMC techniques to assess the significance of detected oscillations robustly. Although these Bayesian methods offer a statistically robust framework, their computational complexity can be prohibitive for the large datasets typical in solar imaging. In our work, we initially employed the MCMC-based approach but found it too slow for our research purposes. To overcome this limitation, we decided to use machine learning. In recent years the field of machine learning has bloomed in popularity in many scientific fields, like in solar physics ( Asensio Ramos et al. 2023 , see) , for the benefits it provides in computational efficiency, and the ability to process enormous amounts of data. For this reason, we developed a deep learning model replicating the MCMC method’s predictive capabilities, substantially improving computational efficiency without compromising accuracy. The

Read the complete text