Overview
The Ministry of Science & Technology has supported a project that uses Artificial Intelligence to digitise a century‑old collection of hand‑drawn solar observations from the Kodaikanal Solar Observatory (KoSO). By converting these sketches into machine‑readable data, scientists can study long‑term variations in solar magnetic activity, which affect satellite operations, power grids and navigation on Earth.
Key Developments
- Researchers led by Dibya Kirti Mishra of ARIES applied a supervised U‑Net model to locate the solar disk and detect solar plage in 100‑year suncharts.
- The model processed drawings covering Solar Cycles 15 to 23 (1916‑2007), producing a continuous “butterfly diagram” that maps the latitude migration of plage areas over each cycle.
- Plage areas derived from the hand‑drawn charts showed strong agreement with measurements from KoSO’s Ca II K full‑disk photographs, confirming the reliability of the digitised data.
Important Facts
- KoSO’s suncharts span from 1904 to 2022, recording sunspots, plages, filaments and prominences on a standard grid.
- Traditional digitisation was hampered by variations in drawing style, paper ageing and scan quality.
- The AI workflow involved two steps: (i) automatic detection of the Sun’s disk (center, radius, tilt) and (ii) segmentation of plage regions using the trained U‑Net.
- The resulting dataset fills gaps in existing solar‑activity records, enabling better reconstruction of the Sun’s energy output over a century.
Exam Relevance
Understanding solar magnetic cycles is part of space‑weather studies, a recurring theme in GS3 (Science & Technology). The project illustrates how machine learning can modernise historical scientific archives, a point of interest for questions on technology adoption and data‑driven policy making. Moreover, the collaboration between Indian institutes (ARIES, IIA, IIST) and an international partner (Southwest Research Institute, USA) showcases the role of international scientific collaboration in advancing national capabilities.
Way Forward
- Extend the AI pipeline to other historic solar archives (e.g., Mt. Wilson, Greenwich) to build a global, century‑long magnetic‑activity database.
- Integrate the digitised plage data with modern satellite observations for improved forecasting of solar storms.
- Encourage policy support for AI‑based preservation of scientific heritage, linking it to national security and infrastructure resilience.