Connecting Solar Physics Past to Its Machine Learning Future

将过去的太阳物理学与其机器学习的未来联系起来

• 批准号: 2035710
• 负责人: Geoffrey Crowley
• 金额: $ 52.55万
• 依托单位: ATMOSPHERIC & SPACE TECHNOLOGY RESEARCH ASSOCIATES, L.L.C.
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-04-15 至 2025-03-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2035710&HistoricalAwards=false
• 关键词: Connecting; Solar; Physics; Past; Its

People have been studying the Sun for hundreds of years and drawing maps of what they see, the earliest dating back almost 900 years. These hand-drawn maps of the solar surface have been invaluable for modern-day physicists to understand how the Sun changes over the decades. Between 1964 and 2009, Dr. Patrick McIntosh made his own carefully constructed maps of the Sun, documenting many interesting solar features and magnetic fields. However, since the last maps were made in 2009, McIntosh’s archive has languished and has not been updated. The purpose of this 3-year project is to use the latest space- and ground-based images of the Sun to reproduce the McIntosh’s techniques, update the archive through today, and develop tools for current machine learning algorithms to produce these maps going into the future.All solar magnetic fields are fundamentally driven by an internal dynamo. To study this pervasive magnetic field, scientists rely on indirect measurements of visible plasma features at the surface that are produced by the evolving magnetic field. Large-scale atmospheric features, such as filaments, active regions, sunspots, polarity boundaries, and coronal holes are distinct signatures of the ever-changing dynamo. These visible features were captured by Dr. Patrick McIntosh from 1964 through 2009 in hand-drawn maps of the Sun. This 3-year project aims to preserve multiple decades of accumulated specialized knowledge of the Sun, thus opening up new avenues for studies of the solar cycle by linking the McIntosh-identified solar features to modern images of the Sun through reproducing his techniques with digital tools. The result of this effort is a new archive of solar maps to effectively extend the existing catalogue to cover solar cycle 24. The catalogue will be open for public download and well documented for supervised machine learning tools to take advantage of this unique labeled data representing over 55 years of solar measurements. The research and EPO agenda of this project supports the Strategic Goals of the AGS Division in discovery, learning, diversity, and interdisciplinary research.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

数百年来，人们一直在研究太阳，并绘制他们所看到的地图，最早的可以追溯到近900年前。 这些手绘的太阳表面地图对于现代物理学家了解太阳在过去几十年中的变化非常宝贵。 在1964年到2009年之间，帕特里克麦金托什（Patrick McIntosh）博士制作了他自己精心绘制的太阳地图，记录了许多有趣的太阳特征和磁场。 然而，自从2009年制作了最后一张地图以来，麦金托什的档案一直萎靡不振，没有更新。 这个为期3年的项目的目的是使用最新的空间和地面太阳图像来重现麦金托什的技术，更新今天的档案，并为当前的机器学习算法开发工具，以产生未来的这些地图。所有的太阳磁场基本上都是由内部发电机驱动的。 为了研究这种普遍存在的磁场，科学家们依赖于对表面可见等离子体特征的间接测量，这些特征是由不断变化的磁场产生的。 大尺度的大气特征，如暗条、活动区、太阳黑子、极性边界和冕洞，都是不断变化的发电机的独特特征。 这些可见的特征是由帕特里克麦金托什博士从1964年到2009年在手绘的太阳地图中捕捉到的。 这个为期三年的项目旨在保存几十年积累的太阳专门知识，从而通过利用数字化工具复制麦金托什的技术，将他确定的太阳特征与现代太阳图像联系起来，为太阳周期研究开辟新的途径。 这一努力的结果是建立了一个新的太阳地图档案，有效地扩大了现有的目录，以涵盖太阳活动周期24。 该目录将开放供公众下载，并为监督机器学习工具提供良好的文档记录，以利用代表超过55年太阳测量的独特标记数据。 该项目的研究和EPO议程支持AGS部门在发现、学习、多样性和跨学科研究方面的战略目标。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。