Motivation: How is magnetic energy transported across the photosphere and the chromosphere? This energy heats the chromosphere and corona, and controls variety of atmospheric phenomena (e.g., fibrils, spicules, jets, surges, flux ropes). The rate of magnetic energy input, or Poynting flux, depends on both magnetic fields and plasma velocities (flows). While derivation of magnetic fields from spectropolarimetric inversions has been heavily studied, much less effort has been devoted to methods of inferring flows. Therefore, the development and validation of observation-based flow reconstruction techniques is critical for understanding the input of magnetic energy into solar atmosphere and its role in driving a variety of solar phenomena.
Methods: Various methods have been developed to estimate flows from observations in different areas of solar physics. (1) “Optical flow” methods rely on algorithms inspired by computer vision. As input, they use a time series of satellite or ground-based observations of the solar photosphere and above, typically tracking flows and motion of small magnetic elements in the Quiet Sun. (2) Physics-based methods calculate flows that satisfy some governing equation, such as the induction equation for the plasma velocity or electric-field vector. As inputs, these methods use Doppler velocities and vector magnetic field observations, typically in the active Sun. (3) Finally, supervised machine-learning methods, some using deep learning, are used in conjunction with numerical simulations of the solar surface and atmosphere to learn a mapping function from imagery or spectropolarimetric inversions to the depth-dependent plasma velocity (i.e., neural networks approximating flows in simulations). More recently, industry-inspired unsupervised learning methods have been suggested to measure optical flows from solar observations.
Proposal: We are a multidisciplinary International Team at ISSI to bring together experts on various flow inversion techniques, including observations, theory, simulations, computer vision and deep learning. Our objective will be to develop, validate and improve flow-tracking methods for a wide range of use cases in solar physics, including simulations with known ground truths and observations. Our collaboration, fostered by two meetings at ISSI-Bern, will yield improved understanding of the plasma flows and magnetic energy flux that permeate the photosphere and chromosphere, catalyzing progress toward the ultimate goal of understanding the input of magnetic energy from the Sun’s interior into the corona and heliosphere. This effort will intensity interactions between the observational and theoretical communities to provide new insights into our understanding of flows, and the nature of magnetic energy input into the solar atmosphere. We will also prepare an open-source community toolkit applicable for existing and future datasets.