Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere

Details Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere

May 2011

adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2011spd....42.2217n&link_type=abstract

American Astronomical Society, SPD meeting #42, #22.17; Bulletin of the American Astronomical Society, Vol. 43, 2011

Statistics

Computation

Scientific paper

We applied principal component analysis to 8-10 hour sequences of chromospheric Hα images of selected solar active regions as observed by the U. S. Air Force Improved Solar Optical Observation Network (ISOON) telescope at Sacramento Peak, NM at one-minute intervals. A covariance matrix of all combinations of image time pairs was computed from the picture element Hα intensities from each image sequence, and eigenvalues and eigenvectors were computed. Computation of explained variance from the eigenvalues indicated that 99.9% of the characteristics were represented by the first 50 eigenvectors or so. The leading eigenvectors were matched at each image time with a flare category indicator deduced from coincident active region area-average Hα intensity and 1-8 Å GOES X-ray flux measurements at one-minute intervals. Multivariate discriminant analysis (MVDA) was applied to the eigenvector elements and flaring indicators from a training set of image sequences to compute a vector of coefficients whose linear combination with the eigenvector predictors maximizes the distinction among flaring groups. The discriminant function, computed from the dot product of the coefficients with the eigenvectors of an independent ISOON case, is used to compute the probability of each flaring group at each measurement time. Flare category indicators from independent case times are used to assess the flaring probability estimates. Multivariate logistic regression (MVLR) was also applied to the subset of eigenvectors in the training data, and the derived coefficients were also dotted with the independent eigenvalues to generate a separate flaring probability estimate. A total of 47 ISOON image sequences were available for training and flaring probability estimation. Preliminary results showed that the MVLR was competitive with MVDA in skill of diagnosing flaring probability. If this approach shows promise in a diagnostic mode, efforts will be made to extend it to short-term (1-3 hour) prognostic mode.

Affiliated with

Also associated with

No associations

Rating

Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.

If you have personal experience with Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Estimating Flaring Probability from High-Cadence Images of the Solar Chromosphere will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFWR-SCP-O-1205358