Inferring depth-dependent plasma motions from surface observations using the DeepVel neural network

Tremblay, Benoit; Cossette, Jean-François; Kazachenko, Maria D.; Charbonneau, Paul; Vincent, Alain

doi:10.1051/swsc/2020073

Cited by 3 publications

(3 citation statements)

References 45 publications

(49 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our work, we pick σ = 10, as it produces the strongest correlation. Following the analyses in Schrijver et al (2006) and Tremblay et al (2021), we consider three other correlation metrics between the reference STAGGER velocities and the velocities obtained from inversions: the spatially averaged relative error…”

Section: Transverse Velocity Inversion Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Quantifying Poynting Flux in the Quiet Sun Photosphere

Tilipman,

Kazachenko,

Tremblay

et al. 2023

ApJ

Self Cite

View full text Add to dashboard Cite

Poynting flux is the flux of magnetic energy, which is responsible for chromospheric and coronal heating in the solar atmosphere. It is defined as a cross product of the electric and magnetic fields, and in ideal MHD conditions it can be expressed in terms of the magnetic field and plasma velocity. Poynting flux has been computed for active regions and plages, but estimating it in the quiet Sun (QS) remains challenging due to resolution effects and polarimetric noise. However, with the upcoming DKIST capabilities, such estimations will become more feasible than ever before. Here, we study QS Poynting flux in SUNRISE/IMaX observations and MURaM simulations. We explore two methods for inferring transverse velocities from observations—FLCT and a neural network–based method DeepVel—and show DeepVel to be the more suitable method in the context of small-scale QS flows. We investigate the effect of azimuthal ambiguity on Poynting flux estimates, and we describe a new method for azimuth disambiguation. Finally, we use two methods for obtaining the electric field. The first method relies on an idealized Ohm’s law, whereas the second is a state-of-the-art inductive electric field inversion method PDFI_SS. We compare the resulting Poynting flux values with theoretical estimates for chromospheric and coronal energy losses and find that some of the Poynting flux estimates are sufficient to match the losses. Using MURaM simulations, we show that photospheric Poynting fluxes vary significantly with optical depth, and that there is an observational bias that results in underestimated Poynting fluxes due to an unaccounted shear term contribution.

show abstract

Section: Transverse Velocity Inversion Methodsmentioning

confidence: 99%

“…) inTremblay et al (2021). For FLCT, the values of these metrics are (E rel = 1.09, C = 0.35, A = 0.21).…”

mentioning

confidence: 99%

Quantifying Poynting Flux in the Quiet Sun Photosphere

Tilipman,

Kazachenko,

Tremblay

et al. 2023

ApJ

Self Cite

View full text Add to dashboard Cite

show abstract

“…DeepVel can estimate the horizontal velocity at various heights in the solar atmosphere without averaging. Tremblay et al (2021a) showed that the Pearson linear correlation between the estimation and the answer was approximately 0.8. The correlation increases when the horizon- tal velocity fields are averaged over several granular lifetimes (Tremblay et al 2018); the increment of the accuracy by taking the average shows the same trend as the LCT.…”

Section: Introductionmentioning

confidence: 98%

Multi-scale deep learning for estimating horizontal velocity fields on the solar surface

Ishikawa

Nakata

Katsukawa

et al. 2022

A&A

View full text Add to dashboard Cite

Context. The dynamics in the photosphere is governed by the multi-scale turbulent convection termed as granulation and supergranulation. It is important to derive three-dimensional velocity vectors to understand the nature of the turbulent convection and to evaluate the vertical Poynting flux toward the upper atmosphere. The line-of-sight component of the velocity can be obtained by observing the Doppler shifts. However, it is difficult to obtain the velocity component perpendicular to the line of sight, which corresponds to the horizontal velocity in disk center observations. Aims. We present a new method based on a deep neural network that can estimate the horizontal velocity from the spatial and temporal variations of the intensity and vertical velocity. We suggest a new measure for examining the performance of the method. Methods. We developed a convolutional neural network model with a multi-scale deep learning architecture. The method consists of multiple convolutional kernels with various sizes of receptive fields, and performs convolution for spatial and temporal axes. The network is trained with data from three different numerical simulations of turbulent convection. Furthermore, we introduced a novel coherence spectrum to assess the horizontal velocity fields that were derived for each spatial scale. Results. The multi-scale deep learning method successfully predicts the horizontal velocities for each convection simulation in terms of the global correlation coefficient, which is often used to evaluate the prediction accuracy of the methods. The coherence spectrum reveals the strong dependence of the correlation coefficients on the spatial scales. Although the coherence spectra are higher than 0.9 for large-scale structures, they drastically decrease to less than 0.3 for small-scale structures, wherein the global correlation coefficient indicates a high value of approximately 0.95. By comparing the results of the three convection simulations, we determined that this decrease in the coherence spectrum occurs around the energy injection scales, which are characterized by the peak of the power spectra of the vertical velocities. Conclusions. The accuracy for the small-scale structures is not guaranteed solely by the global correlation coefficient. To improve the accuracy on small scales, it is important to improve the loss function for enhancing the small-scale structures and to utilize other physical quantities related to the nonlinear cascade of convective eddies as input data.

show abstract

Intensification of magnetic field in merging magnetic flux tubes driven by supergranular vortical flows

Chian

Rempel

Silva

et al. 2022

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

The spatiotemporal dynamics of vorticity and magnetic field in the region of a photospheric vortex at a supergranular junction of the quiet Sun is studied, using Hinode’s continuum intensity images and longitudinal magnetograms. We show that in a 30 min interval during the vortex lifetime, the magnetic field is intensified at the centres of two merging magnetic flux tubes trapped inside the vortex boundary. Moreover, we show that the electric current density is intensified at the interface boundary layers of merging tubes, resulting from strong vortical downflows in a supergranular vertex. Evidence of Lagrangian chaos and vortex stretching in the photospheric plasma turbulence responsible for driving the intensification of magnetic fields is analysed. In particular, we report the first solar observation of the intensification of electromagnetic energy flux resulting from the merger of magnetic flux tubes.

show abstract

Inferring depth-dependent plasma motions from surface observations using the DeepVel neural network

Cited by 3 publications

References 45 publications

Quantifying Poynting Flux in the Quiet Sun Photosphere

Quantifying Poynting Flux in the Quiet Sun Photosphere

Multi-scale deep learning for estimating horizontal velocity fields on the solar surface

Intensification of magnetic field in merging magnetic flux tubes driven by supergranular vortical flows

Contact Info

Product

Resources

About