An AI-powered Surface Flux Transport model to measure high-resolution velocity fields and forecast magnetic flux emergence

Speaker: Andrés Muñoz-Jaramillo

Dec 9, 2025 11:00 PST

Solar magnetism is the main driver of solar and interplanetary variability. The evolution of the solar magnetic field is driven largely by the emergence of complex magnetic regions into the solar photosphere and corona, and their subsequent evolution as they interact with plasma flows and existing magnetic fields. Surface flux transport (SFT) models are the main tool we use to understand the evolution of surface magnetism; they work by driving the evolution of surface magnetic fields through the prescription of surface velocity fields and flux emergence. Currently, the prescription of flows and fields in SFT models is based on independent studies of solar surface activity. However, Artificial Intelligence learning frameworks present a unique opportunity to develop an SFT that can combine the simulation of transport with the exercise of measuring and estimating surface flows and emerging fields. Here we present the ARCADE (Active Region Characterization and Analysis of Evolution) model: A hybrid model that uses data driven AI estimation of flux emergence and velocity fields and integrates the surface flux transport equation using PyTorch. We discuss the power of combining numerical integration into a PyTorch framework, how our model performs in comparison to state-of-the-art SFT models, provide an overview of estimated velocity fields and flux emergence patterns, and discuss how it can be used to expand the toolbox that we use to understand the Consequences Of Fields and Flows in the Interior and Exterior of the Sun.

Authors: Andrés Muñoz-Jaramillo1, Nina Bonaventura2, Plinio Guzmán3, Nishu Karna4, Katherine Keegan5, Bibhuti Kumar Jha1, Shea Hess-Webber6, & Spiridon Kasapis7

  1. Southwest Research Institute, Boulder, CO
  2. Steward Observatory, University of Arizona, Tucson, AZ
  3. Fused.io, pgzmnk@gmail.com
  4. Center for Astrophysics | Harvard and Smithsonian, Cambridge, MA
  5. Department of Mathematics, Emory University, Atlanta, GA
  6. Hansen Experimental Physics Lab, Stanford University, CA
  7. Space Physics Group, Department of Astrophysical Sciences, Princeton University, NJ