JpGU-AGU Joint Meeting 2020

Presentation information

[E] Poster

P (Space and Planetary Sciences ) » P-EM Solar-Terrestrial Sciences, Space Electromagnetism & Space Environment

[P-EM14] Frontiers in solar physics

convener:Shinsuke Imada(Institute for Space-Earth Environmental Research, Nagoya University), Alphonse Sterling(NASA/MSFC), Takaaki Yokoyama(School of Science, University of Tokyo), Toshifumi Shimizu(Institute of Space and Astronautical Science, JAXA)

[PEM14-P02] Image analysis of sunspot drawings in Dalton minimum(the early 1800s) and reproduction of magnetograms from sunspot drawings by using machine learning.

*SHOMA UNEME1, Shinsuke Imada1, Hisashi hayakawa2, Lee Harim3, Moon Yongjae3, Park Eunsu3, Yoshizumi Miyoshi1 (1.nagoya university institutive for Space Earth for enviornmental research, 2.Osaka University School of Letters , 3.School of Space Research, Kyung Hee University)

Keywords:Solar activity, Dalton minimum, machine learning, Sunspot drawings

Solar activity varies periodically in 11 years. Because the solar activity is one of the main
origins of the variability of the solar-terrestrial environment, it is important to predict the
solar activity especially in the category of space weather study. It is known that the polar
magnetic field at the solar minimum is closely correlated with the solar activity at the
next solar activity. This correlation was confirmed by observing the current sun precisely.
On the other hand, it is not clear whether there was a similar correlation in the past sun.
Therefore, the aim of this study is to analyze the past sunspot drawing images to verify
whether the polar field value at the solar minimum is also good correlation with the next
solar activity. Especially, we focused on the Dalton minimum when sunspots were little
in the early 1800s. We extracted latitude and longitude of sunspot from drawings in the
early 1800 's. Furthermore, we generated magnetograms from sunspot drawings by using
cGAN(Conditional Generative Adversarial Nets), one of the popular machine learning
methods, and will discuss the validity of these magnetograms. In future, we will calculate
the polar magnetic field in the Dalton minimum from these generated magnetograms by
surface magnetic flux transport model and compare it with the correlation.