Diagnosis of Solar Flare Probability from Chromosphere Image Sequences

Abstract

We used optical observations of the solar chromosphere in the diagnosis of flare probability at the observation time. We derived eigenvectors from sequences of hydrogen-alpha (H-alpha) images of sub-regions of selected solar active regions at one-minute intervals. A degree of flaring category was set from H-alpha intensity and 1-8 Angstrom x-ray flux. Leading eigenvector elements were the predictors, and flaring category the predictand, in employing multivariate discriminant analysis (MVDA) on a training set of image sequences. We applied resulting discriminant vectors to image sequence eigenvector elements to diagnose flaring category probability. Metrics of comparison with the specified flaring category determined diagnosis skill. Binary no-flare/flare predictand categories set with a specified H-alpha rise produced too many flaring diagnoses. Setting multiple flaring categories from H-alpha magnitude produced a large positive category bias. Assigning flaring categories according to an H-alpha eigenvector pattern - x-ray flare intensity association improved the category diagnosis performance. The latter method will be further refined and applied to additional ISOON image sequences.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 30, 2011
Accession Number
ADA554688

Entities

People

  • Donald C. Norquist
  • K. S. Balasubramaniam

Organizations

  • Air Force Research Laboratory

Tags

Communities of Interest

  • Energy and Power Technologies
  • Materials and Manufacturing Processes
  • Space

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Chromosphere
  • Data Science
  • Discriminant Analysis
  • Eigenvectors
  • Elements
  • Information Science
  • Intensity
  • Intervals
  • Observation
  • Probability
  • Solar Cycle
  • Solar Disturbances
  • Solar Flares
  • Statistical Algorithms
  • X Rays

Fields of Study

  • Physics

Readers

  • Neural Network Machine Learning.
  • Regression Analysis.
  • Solar Physics