NSWP: Automated Monitoring and Forecasting of Space Weather using Artificial Intelligence Techniques

  • Jing, Ju J. (PI)
  • Shih, Frank (CoPI)
  • Wang, Haimin H. (CoPI)

Project: Research project

Project Details


The proposing team will use pattern recognition techniques to predict the occurrence of solar flares, developing a new tool known as the Support Vector Machine (SVM). The proposers intend to develop tools to detect new magnetic flux emergence on the Sun, using circular harmonic component decomposition as a filter for an artificial intelligence classifier. This technique will characterize new flux emergence and establish the probabilities for active regions to become flare productive. They will implement a solar flare detection and characterization algorithm, including a classification scheme using the SVM, active region growth, and edge enhancement techniques. The algorithm will detect flare ribbon separations and help determine the electric currents in magnetic reconnection regions. The PI will also study a large number of CMEs using a characterization routine to establish the relationship between CME speed and magnetic reconnection rate. This will allow the prediction of CME kinetics based on real- time monitoring of magnetic reconnection.

This effort will enhance our understanding and prediction of processes affecting solar activity and the propagation of resulting solar effects to the Earth via the solar wind. The work is inherently interdisciplinary, involving cutting-edge solar physics and computer science research. The techniques developed here also have potential utility for medical imaging, terrestrial weather forecasting, and pattern recognition for moving targets relevant to military applications. This proposal's education and training component involves the support of a newly graduated post-doctoral researcher and a graduate student.

Effective start/end date8/1/077/31/10


  • National Science Foundation: $151,415.00


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.