Документ взят из кэша поисковой машины. Адрес оригинального документа : http://hea-www.harvard.edu/AstroStat/Stat310_1011/haiku_jpb_20100927.pdf Дата изменения: Mon Sep 27 08:10:51 2010 Дата индексирования: Tue Oct 2 06:03:37 2012 Кодировка:

Power-laws and Solar Flares
Jennifer Posson-Brown, Paolo Grigis & Vinay Kashyap (CfA)

Peak intensity of regular solar flares follows power-law distribution with alpha=1.8. Do weaker flares have the same distribution?

Sunday, September 26, 2010


Data from XRT on Hinode

Difficulty: Detecting flare events in the presence of variable background
Sunday, September 26, 2010


Flare Detection Strategies
Currently we use 2 detection methods based on * time derivatives * segmenting based on significant minima and use simulated lightcurves to optimize parameter values. We want to try Alex Blocker's wavelet method!

Sunday, September 26, 2010


black = observed, blue = best match

Is this alpha=1.8 or alpha=2.5? Monte Carlo simulations to find the best match to count rate distributions, BUT: can find a good match regardless of alpha!

Sunday, September 26, 2010


Can we use comparison with simulated datasets to constrain alpha?
(Jiashun Jin, Purdue University)

Define range of acceptable Chi-square values and see what % of simulations are in that range for given alpha?

Sunday, September 26, 2010


smooth LC w/ fixed-width Gaussian; look at where derivative crosses certain thresholds to find flare start and end (parameters: smoothing width, 3 thresholds, threshold normalization factor)

Sunday, September 26, 2010


Loess smoothing on LC; define segments as intervals between consecutive minima, merge segments if adjacent extrema statistically indistinguishable (parameters: smallest scale for smoothing, S/N threshold for merging adjacent segments)

Sunday, September 26, 2010