Title and Abstract

Title:  SPI3S: SPectral Irradiance of a 3d Sun

Abstract:   The Sun influences life on Earth and impacts our technology in space. For example, the Extreme Ultraviolet light (or EUV) that the Sun emits affects Earth’s atmosphere in a way that degrades the orbit of satellites. The Sun is also the only star in the universe that can be spatially resolved. This makes the Sun an important template against which our broader understanding of stars is validated.

Over the last 15 years, at most three satellites observed the Sun in EUV at any given time. Additionally, these satellites have been tied to the ecliptic plane, thus orbiting and imaging the Sun from its equator and incapable of observing the solar poles directly. However, a complete image of the Sun is required to forecast EUV irradiance to protect our assets and to relate the Sun to other stars in the universe that are otherwise observed as distant points in the sky with unknown inclination angles.

The SPI3S pipeline (SPectral Irradiance of a 3d Sun) uses EUV images of the Sun taken from multiple viewpoints to train neural networks called Sun Neural Radiance Fields (SuNeRFs) and create 3D representations of the solar atmosphere in EUV. It also uses EUV images and EUV spectra to train a neural network to predict, from novel EUV images, the irradiance at each point in the solar system.

In addition to generating novel views of the Sun beyond what existing satellites can achieve (including viewing the solar poles), our end-to-end SPI3S pipeline enables the short-term forecast of EUV radiation impacting Earth and, for the first time, the direct comparison of solar EUV spectra with that of any observable star. This paves the way for using EUV spectra to better understand solar and stellar magnetism.  It is also an example of how novel machine learning techniques can be used to significantly enhance observational capabilities by the creation of virtual instruments.