3rd August 2006

Well, it's all calmed down a bit and the program is in its final stages.  I have not had to display anything on this page to share with supervisors etc., but if you fancy wading through a folder of images most of my graphs and images are available here:

Folder of Figures:  http://solar.physics.montana.edu/home/www/reu/2006/arussell/figs/

Alternatively, I strongly suggest taking a look at my end of program report.  It is available in two forms: a detailed report (similar to a paper); or as slides used for my end of year presentation (treat as a summary, but perhaps best for those with some familiarity).

Detailed Report:      http://solar.physics.montana.edu/home/www/reu/2006/arussell/report.html

Presentation Slides: http://solar.physics.montana.edu/home/www/reu/2006/arussell/pres.html

Enjoy!


26th July 2006

A few days ago I investigated the effect of the flux cutoff on s. Decided to post the figure today.


25th July 2006

Today I gave quite a bit of consideration to the previous days results,looking for interpretation. As part of this I calculated a new "distribution parameter" (a new, more appropriate name for what I used to call "measures of fragmentation"). This distribution parameter is the variance of ln(flux_of_partition), calculated for each active region. I now believe that the correlation between s^2 and the average flux per partition is a result of method. The correlation between the number of fragments and the active region area (total area of all concentrations in the active region) is a physical result. I interpret it as showing there is a favoured area for each fragment, resulting from the equilibrium between fragmentation and coalenscence of flux tubes. The distance across this area is approximately half the distance across a convective granule.

I also computed partial correlations involving flaring. Both distributed parameters showed a negative partial correlation with flaring (flux held constant). The partial correlation between number of fragments and flaring, with flux held constant, was 0.224545. A greater correlation, of 0.574510, was found between the numnber of fragments and log10(flaring) with flux held constant. For comparison, the partial correlation between flux and EF with the number of fragments held constant was 0.239609 - surprisingly low. However, the partial correlation between flux and EF rose to 0.699688 when the variance was held constant instead of the number of fragments.


Distribution Parameters

Pearson coeff.= 0.793453
Spearman coeff.= 0.755582
Spearman sig.= 4.63726e-12

Number of Fragments and Total Area

Pearson coeff.= 0.972860
Spearman coeff.= 0.967049
Spearman sig.= 1.47606e-35

Variance and Flaring

Pearson coeff.= 0.0745913
Spearman coeff.= 0.301572
Spearman sig.= 0.0202771

Pearson coeff.= 0.118509
Spearman coeff.= 0.261421
Spearman sig.= 0.135326

Note: For non-zero flaring only

Variance and Other Parameters

Pearson coeff.= -0.0369664
Spearman coeff.= -0.0160151
Spearman sig.= 0.904175

Pearson coeff.= 0.285858
Spearman coeff.= 0.284687
Spearman sig.= 0.0288612

Pearson coeff.= 0.845199
Spearman coeff.= 0.753536
Spearman sig.= 5.70293e-12

This becomes even stronger for a consideration of partitons with flux greater than the cutoff (Pearson coeff. = 0.936209 ). I firmly believe that this is a result of method, rather than any physical result.

Pearson coeff.= 0.388729
Spearman coeff.= 0.388311
Spearman sig.= 0.00237300


24th July 2006

Summary: I have developed software to identify concentrations of magnetic flux in magnetograms of solar active regions. This software has been applied to a database of cropped full disk magnetograms. A measure of fragmentation was then sought. In the figures below, this is s2. s is obtained for each active region by fitting a curve to its cdf, the curve assuming flux is lognormally distributed in each region (the validity of this assumption is assessed from the K-S statistic).


Number of Fragments and Total Flux

Pearson coeff.= 0.887227
Spearman coeff.= 0.920582
Spearman sig.= 6.05764e-25

Flux and Flaring

Pearson coeff.= 0.689946
Spearman coeff.= 0.557178
Spearman sig.= 4.57290e-06

Pearson coeff.= 0.701338
Spearman coeff.= 0.604889
Spearman sig.= 0.000150995

Note: For non-zero flaring only

Pearson coeff.= 0.710127
Spearman coeff.= 0.557178
Spearman sig.= 4.57290e-06

Pearson coeff.= 0.656532
Spearman coeff.= 0.656532
Spearman sig.= 0.000150995

Note: For non-zero flaring only

Number of Fragments and Flaring

Pearson coeff.= 0.687124
Spearman coeff.= 0.416333
Spearman sig.= 0.00103890

Pearson coeff.= 0.646884
Spearman coeff.= 0.537968
Spearman sig.= 0.00103244

Note: For non-zero flaring only

s2 and Flaring

Pearson coeff.= -0.00731173
Spearman coeff.= 0.160332
Spearman sig.= 0.225112

Pearson coeff.= 0.0359645
Spearman coeff.= 0.129412
Spearman sig.= 0.465730

Note: For non-zero flaring only

s2 and Other Parameters

Pearson coeff.= -0.158929
Spearman coeff.= -0.207292
Spearman sig.= 0.115179

Pearson coeff.= 0.635276
Spearman coeff.= 0.599123
Spearman sig.= 5.35453e-07

Pearson coeff.= 0.0777658
Spearman coeff.= 0.0492694
Spearman sig.= 0.710956

Welcome About Me Research Interests Research Log IDL Tools Photos Links