Jada's Daily REU Log

Home

Jiong's Daily Log

2.august.2006, Wed

Today and yesterday I continued working on my presentation. I also did a rough run-through of the presentation. I now have quite a solid idea of what I will talk about and what I should exclude.

Last night I also began re-reading all of the papers that Jiong gave me. My understanding of much of the jargon and processes really has come a long way. Re-reading these papers is really helping me to pull it all together and better understand my contribution to a larger body of research.



31.july.2006, Mon

Today I mostly worked on my presentation template and layed out the presentation pages. I've also got an idea of the images and movies that I will use, which I will begin to load into the presentation tomorrow.



30.july.2006, Sun

As I said yesterday, the RHESSI sources don't appear to move much. I made an animated gif to see how the TRACE ribbons evolve over the same period as our RHESSI data. The ribbons expand very slowly during this time. If you watch a movie of the entire flare, you will see that only in the beginning of ribbon formation and expansion do the ribbons expand quickly - but this is only due to the lower cadence of the exposures during the beginning of the data. This flare has a very long decay phase and it seems that there is ribbon expansion well into the "decay".



29.july.2006, Sat

Jiong will be gone to the SHINE Workshop this week (some of my results will be included in her presentation). I will mostly be working on my presentation, but this weekend I'm spending a bit of time investigating the RHESSI footpoint/TRACE flare ribbon relationship. Jiong wrote a program to contour several good RHESSI images over one image to get an idea of the morphology of the sources. Interestingly, the sources (25-100 keV) don't appear to move much at all over the approximately 30 second time frame. I'm now looking into how the flare ribbons move during this time period and finding out what the magnetic field is like where these sources appear.

I modified Jiong's program (analyze_hessi.pro) to contour the good RHESSI frames over MDI (analyze_hessi_on_mdi.pro). I scaled the MDI image to +/- 300 Gauss and it appears that the RHESSI sources do appear where the magnetic field is stongest - although I have questions about the alignment... (Jiong can you double check? I think that I should use the original (not fine tune aligned) MDI image because it contains the original position of MDI which RHESSI will contour to - is this correct?).


IMG 1 MDI fine aligned with TRACE
IMG 2 RHESSI contours on fine aligned MDI
IMG 3 MDI NOT fine aligned with TRACE
IMG 4 RHESSI on non-fine aligned TRACE

2 MDIs Fine tune and non-fine tune next to each other, with contours



28.july.2006, Fri

Jiong and I talked again about where our uncertainties come from, how we account for them in our analyses and how we calculate them. We also discussed how MC are measured and why the kinetic energy is relevant to the reconnection flux. We discussed MDI saturation and I decided to, in the future, look at the Kitt Peak magnetograms to compare them to the MDI magnetograms that I used.

At the end of the day I used plot_map to make quick overlays of RHESSI, TRACE and MDI. I found a good RHESSI image from very shortly after the time of the TRACE reference frame that I used. The RHESSI image that I used was 25-100 keV. The only one that really looks like anything is the TRACE overlay on RHESSI. This second one may show the flare ribbons over the footpoints more clearly.



27.july.2006, Thur

I made a few RHESSI images and then made a huge image cube. Since Emily and Crystal showed their RHESSI images at our first meeting, I've been curious about how the footpoints relate to the flare ribbons. There was only RHESSI data for the last flare I analyzed.



26.july.2006, Wed

See plots.



25.july.2006, Tue

I updated the event table to include positive and negative flux, flux ratio and flare location (this one because I was curious if the leading flux and the flux ratio might have a relationship). I also made a structure with most of the data from the event table. Jiong and I went through plotting some of the data and found that there is a nearly linear relationship between the reconnection flux and the kinetic energy of the CME. Tomorrow I will get more familiar with plotting data in PostScript.



24.july.2006, Mon

Today I am beginning to put all of the data together - my events with the previous events which Jiong analyzed. I made an event table which lists all of the relevant information. Next, I will put most of this data together in a structure so I can make plots in IDL.



21.july.2006, Fri

This morning I reanalyzed the 2005.May.05 event. The afternoon was spent preparing the Yellowstone solar talk for tomorrow.



20.july.2006, Thur

This morning Jiong found another problem which likely contributed to the low and imbalanced flux of 050513. I apparently created sub-MDI maps and images with a low magnetic strength threshold - I use these low thresholds when viewing MDI and looking for weaker field structures, but they need to be set back to a higher threshold for the analysis. I've checked all of the sub-MDI images for the other flares and I saved all of them correctly - they all have high thresholds.

Today I will:

    - download TRACE data for at least an hour and a half after the flare end time
    - remap and image MDI with the higher threshold
    - xy shift this MDI
    - extrapolate the remapped/imaged MDI
    - process late-flare TRACE data and add to earlier data
    - make changes to the analysis program
    - re-analyze this event

The order of the TRACE processing is important, esp here.

    1. shift for xcen & ycen to same reference frame for first set of data
    2. cut frames to 768x768
    3. shift for solar rotation (this should be done before rebinning so the regions will match since the reference frame is 768x768)
    4. rebin to 384x384
    5. finally, remove the dark current and normalize the exposure time

I've created new sub-MDI maps and images with a higher threshold. Also, the processing of this second set of TRACE data, now from 17:40:19 to 20:22:01, is finished and tomorrow I will join this with the first set of data. Then I will analyze the full set.



19.july.2006, Wed

Last night I finished processing the TRACE data and this morning I created the sub-MDI images. I also chose a cut range (7-15) and the test2 (for integration) results are here

ugh. After running a couple of analyses on this event, the total flux results turned out to be lower than expected and quite imbalanced. It turns out that the flare ribbons still appear to be expanding for nearly another hour and a half. H-alpha images which Jiong has in her paper which includes this event revealed this to us. So I will look for and download more TRACE data to include in the re-analysis. Also, the integration was higher than necessary and this will also be changed for the re-analysis.



18.july.2006, Tue

Worked more on the 050513 TRACE data. Also, I reanalyzed the 2004.jan.19 flare with Jiong's revised program. Apparently, by making a smaller region to analyze the background/flare pixel contrast becomes more sensitive so the flux increased, as seen on the strange results page. The new results look more reasonable and can be seen on results page for 2004.01.19.



17.july.2006, Mon

Worked on TRACE data today - removed some 1024x1024 frames, resized these frames to 768 to match the others. The FOV was good and all centers were the same, thankfully.



14.july.2006, Fri

Yesterday, when I was putting together the results page for the 2004.Jan.19 flare, I noticed that the MDI image with the TRACE overlay of the prime integration and cut did not look right. The area of the flare ribbons seemed much to small. Jiong confirmed this and so I will do at least one more analysis until the area of the ribbons looks appropriate.

I played with the cuts and integrations and found some strange results. For this flare I had to cut the region of TRACE and the two MDI images so I could eliminate some of the bright spots so they would not contribute to the reconnection flux. It appeared that sometimes, when I kept the same cut and integration but reduced the region, the total flux would actually increase. Jiong was also surprised by this and is reviewing the data, since the program and my operation of it seem ok.

I've begun to prep the TRACE data for the 2005.May.13 event. The xcen and ycen are the same throughout. But the frame size changes some.

The following frames are 1024x1024:

    - 0:25, I'd like to keep these since they are at the beginning of the flare and show very nice ribbon development
    - every 41 frames between frames 25 and 394, these I will remove
    - 394:413, I'm not sure about these frames, they are mostly cooling but I think there still may be a little bit of ribbon activity.



13.july.2006, Thur

The results page for 2004.Jan.19 is finished.



12.july.2006, Wed

I finished the analysis of 2004.Jan.19 and started putting together a results page. The EIT data was my bane today - which needs to be rotated vertically and horizontally, cropped, contrast adjusted, etc. I plan to finish the results page tomorrow and start on the 2005.May.13 event.



11.july.2006, Tue

This morning I've been working on fine-tune alignments of TRACE and MDI. The TRACE reference frame is at 23:59:06 UT. The chosen MDI reference frame is at 22:23:02 UT - which is appropriate because we want to use an MDI reference frame that occurs before the flare (or at least well before the peak) since the flare ribbons can change the magnetic strength profile of the MDI and interfere during the analysis of the total magnetic flux of the flare ribbon. But after setting points between the reference MDI and TRACE, then shifting the MDI according to the mean x and y coordinates of each point, I'm not quite satisfied with the alignment. After doing setpts 3 times I decided none of them were going to be perfect and chose the best one. Some areas aligned beautifully. In other areas the shift appeared noticeably off. Was this simply due to feature changes over the hour and a half between the MDI and TRACE reference frame times?

I decided to test an MDI frame closer to the time of the TRACE reference frame. The test MDI frame was at 23:59:02 UT. Comparing a TRACE frame taken at 23:59:06 UT, there are noticable flare ribbons, especially in the center of the very large, negative polarity sunspot. I checked the magnetic strength profile of the test MDI image, looking for clean gaussian curves indicating that the profile has not been changed (much, if at all) due to the flare. The horizontal profile was perfect. The verticle profile showed a little dip in the curve of the large sunspot near the middle-left, but overall the magnetic profile of the spot and the rest of the test MDI was smooth. I did setpts twice for the test MDI and found that the alignment may have been only very slightly better than the alignment I did with the original MDI. But both MDI alginments were very good over the region of the flare. To be safe, in case the flare did affect the magnetogram, I will use the earlier MDI image.

I ran through different cuts with choose_cuts.pro and test1 from analyze_flux.pro. There are many bright points which appear throughout the data and last for some time. I cropped the data for the analysis to exclude as much of these bright points as possible so that they would not be included in the total flux.



10.july.2006, Mon

Over the weekend and early this morning I finished the time consuming task of downloading TRACE and MDI. This data, of the same size, good cadence and exposure, is appealing after the last set of flare data. Jiong and I chose a frame range - this is a small flare, C5.5, and the footpoints are more subtle than the others I've analyzed. It's interesting that one of the ribbons is long and expands nicely over the flare time. The other ribbon looks like it's in the umbra of the sunspot, where the magnetic field is strongest, and it is very short and expands slowly. This makes sense to me and I'll have to find the reference that supports the observations that a strong flux expands more slowly. I'm looking forward to looking at the MDI data and analyzing this flare.

So, this MDI data needed to be rotated 180o. This caused problems when overlaying the maps, which mostly worked when it was not rotated, but required that the subframe of MDI be rotated. Also, it was necessary to drot (for solar rotation) in the negative (left) direction. Normally in this process, the MDI image would be rotated positively (to the right) since the reference MDI image is taken before the reference TRACE image.

Since it is late in the day, I will leave the tedious task of fine-tune alignment (aka setpts) until tomorrow morning when I will have fresh eyes.



7.july.2006, Fri

I've requested TRACE and MDI data for the two new events. I'm finishing up putting together images and movies for the 2001.Nov.22 event results page. The eit images are also flipped upside down when I make and mpeg. So now in the command line,

IMAGE = rotate(bytscl(alog(data(*,*,i)>1.)), 7)

- just so i remember to rotate movies in the future.

I spent a lot of time today on EIT images - trying to get movies and images to save with a color that I applied to them which makes the features easier to see. I managed to make a decent movie, although I had very few frames to work with, but the contrast is better than I had it before. The results page for this flare is finished for the most part, although I'd like to add just a bit more.

I'm looking forward to starting the new events next week.



6.july.2006, Thur

Today I'm analyzing the 2001.Nov.22, flare #5580. While choosing cuts, I noticed that the smaller frames (originally 768X768), would not contour for any cut. These frames have poor contrast due to short exposure time (during very high cadence) but this could not have been the problem because some of the large frames also had poor contrast but they would contour.

Jiong quickly realized that these frames would not contour because the extra area (which is black) that surrounds these small frames to compensate for the missing area in relation to the large frames, interferes with the background median. So she showed me how to choose a smaller region to include in the background median calculation.

I made an .mpeg movie of the TRACE data while I was waiting for the analysis to run. It turns out that the programs flips the TRACE data upside down, for some crazy reason. The Bastille Day Flare movie that I made is also upside down - it's not as obvious as this one and so I didn't catch it. I've gone back and rotated the data when making the movies, so now they are correct.

I've finished the analysis for this flare. The results page is in the works.



5.july.2006, Wed

Aligning TRACE & MDI - Here are some things to remember:

    - The reference frame chosen earlier, which all other x.cen and y.cen were shifted to, should be used when doing plot_map because the index and data need to be the same (unshifted).
    - Also, this frame need not be processed.
    - But when doing the fine tune alignment any frame can be used since MDI is now aligned with all of them (after aligning MDI to reference frame).
I did the fine tuning of the alignment for MDI to TRACE. Both the TRACE image and the MDI image were difficult to match points to, so I did two setpoints and chose the second one which aligned very nicely. Then I extrapolated the fine aligned MDI image.

A note for tomorrow: I'll need to choose a bit of a lower cut to include in the analysis. This is because, due to limitations of the TRACE data, the beginning of the flare ribbons are not captured well - basically the data got cut by TRACE near the early part of the flare.



30.june.2006, Fri

The data has been cut to include frames from 23:05:28 to 23:59:42 UT.

Thanks to Jiong, the time cadence has been fixed and she also wrote a program to align the different size frames. Shifting by the x.cen and y.cen only aligns the centers but does not change the field of view. The program she wrote fixed the field of view as well as aligning centers. So now the flare is in the same place throughout the frames.

  • Dark current has been removed
  • TRACE is finally done. Many factors were involved in causing difficulty with shifting all frames to the same place. Some interesting things:

      1. frame 0 is 768x768
      2. frames 1-179 are 512x512
      3. frames 180-241 are 768x768
      4. all 768 frames are good quality
      5. all 512 frames have poor background (due to short exposure time - remember these ones were very high cadence) making choosing a good region to align pixels difficult
    Among the 512 frames there was only a drift due to solar rotation of about 3 pixels. Among the 768 frames there was only a drift of about 3 pixels. But spanning the frames from 1 (512) to 241(768) there was a drift of approximately 6 pixels. Jiong shifted groups of frames and eventually came up with a nice alignment. So now we only have a drift of maybe 2 pixels through all of the frames.



    29.june.2006, Thur

    I've downloaded TRACE data for the new event I will be analyzing, 22 Nov 2001. This event begins around 22:30 UT, with the decay crossing over into 23 Nov. The cadence is very high for most of the flare, about 2-3 seconds, and the frames are mostly 512. There appears to be 3 times where the frame changes to 768 and the cadence during these periods is either about 30 seconds or back to about 3 seconds 23:56:39 UT.



    28.june.2006, Wed

    I've finished the Bastille Day Flare Results page. Out of curiosity, I plan to look at RHESSI data for this event as well -- nope, I won't be as RHESSI was not launched until Feb 2002.



    27.june.2006, Tue

    Today I continued working on the results page for the Bastille Day flare. I removed 3 frames from the TRACE data because they were a little saturated and could have slightly affected the reconnection flux. I re-analyzed the data minus the saturated frames.

    I examined EIT data and found that there is an associated filament eruption. I made a movie of this data and posted that on the site as well.

    I have a little more to do putting together the results page - mostly just tying everything together and giving some explanation about the flux and uncertainties.



    26.june.2006, Mon

    Today I worked on putting together the results for the analysis of the Bastille Day event. I made a movie and an image out of the data that I analyzed. Here is the results page.



    23.june.2006, Fri

    This morning I extrapolated the sub-mdi image and did the fine tune coalignements.

    I had to adjust the data in the final TRACE file because when I originally read the files out, the index listed the frames as 1024x1024. Unfortunately, index only lists the largest frames, which happen to usually be anomalies which need to be removed anyway. The TRACE frames were actually 768x768, so when I rebinned them to 512x512 (thinking they were 1024x1024), the frame included blank space beyond the 384x384 area. I noticed the problem when I did plot_map - the blank area was visible, with the image confined to the 384x384 space. In order to avoid this problem in the future, I can look at the index.nxout and .nyout, which will list all of the frame sizes. Or, when I read the TRACE files, I can read the index only at first, check the frame sizes, set an array to exclude any anomaly frames, then re-read the TRACE files with this array and now including the data.

    I noticed a few things when I ran the TRACE data movie again that I was concerned might cause problems during the analysis. In some of the frames (mostly once the flare begins) the background is smoothed out gray - there is no definition to the background. It seems that this is just because of a short exposure time - there was not enough time for the pixels to pick up light from the background, which is good because otherwise the flare itself would be completly washed out. Also, in some areas of the background there appear tiny black dots - these are likely just pixels that registered lower than the dark current and so remained when I removed the dark current. These artifacts should not have an effect on the flux analysis.

    I go through this everytime I do cuts, so in order to remind myself...the highest cut should still include all of the ribbon. If the range I choose visually is very small, I can simply adjust the range in the program to not inlcude a range of 6 cuts.

    Ran the analysis on the Bastille Day flare (#920). Results page coming soon...



    22.june.2006, Thur

    Continued processing Bastille Day flare:
    -shifted for solar rotation
    -removed dark current and normalized to exposure time
    -TRACE processing is done
    -Mapped and overlayed TRACE and MDI



    21.june.2006, Wed

    Processing the Bastille Day flare - noticed that while choosing a reference frame to shift to, I need to be careful that I don't lose any of the flare in the shift. Choosing a frame in the middle of the flare, making sure the quiet region is stable, should allow the least amount of shifting - a little to the right for frames before the reference and a little to the left for frames after the reference.

    Here is the sequence of steps to take in processing TRACE data:

      1. Select frames based on SEC Event data, GOES X-ray data and visual interpretaion.
      2. Rebin
      3. Shift for pointing by center coordinates (xcen & ycen) if necessary (to check: print, index.xcen, index.ycen to list pointing coordinates). Shift should be based on a frame with a stable quiet region; reference frame should be near the middle of data (see above).
      4. region.pro to pick quiet region from reference frame
      5. xyoff1.pro based on region chosen from reference frame. This finds the shift for solar rotation.
      6. shift_interp to correct for solar rotation
      7. Remove dark current and normalize

      All of this can be done using the prep_trace_template.pro, which has to be customized for each event.


    Today I:

      1. rebinned the TRACE files
      2. chose the frame range (which includes frames about 10 min. before the flare occurs and a few minutes into the cooling of the feet - where there is no more growth and very slight decay in the brightness of the feet can be seen)
      3. checked for pointing shift - pointing was perfect on this flare, so no shift was necessary
      4. removed 2 frames in which the FOV was off (full disk? - still 1024x1024)
      5. correction for solar rotation was necessary, so based on my reference frame I chose a good quiet region to shift to (the first frame I chose (96) turned out to be too washed out to see the background. frame 94 was much better so I used this as a reference frame).
      ...



    20.june.2006, Tue

    This morning I'm working on flare 2350 from the 040725 set. I'm choosing cuts and trying to decide visually which cut range to use. There is quite a difference between using test1 of the analyze_..._flux.pro and choose_cut.pro. This is mostly because test1 takes the median of all of the frames and applies the cut to that. Choose_cut is simply applying the cut to each frame individually. If I take test1 and choose_cut into account, then I should be able to determine a reasonable cut range.

    I've run the analysis on flare 2350 and I've checked with Jiong to make sure the results are reasonable. They look good but there is a discrepancy between the positive and negative flux. This may be due to the TRACE field of view, which may have missed some of the flare. This is also a "weak" flare, although the LASCO CME movies show that it produced a good CME.

    I've requested the Bastille Day (14 July 2000) data, which I will begin processing tomorrow. And tonight I may reorganize my files and directories.

    Also, today I had my first look at the Sun through an H-alpha filter - saw some "little" prominences! Thanks to Ivy and David at the Museum of the Rockies.



    19.june.2006, Mon

    This morning I ran the analyze program on flare 2280 (flare 2) and ran into some problems. The difference in the flux after several cuts and integrations (in test3) was too high. I went through the program to see if I could spot the problem. Some things stood out to me but I wasn't quite ready to mess with them. I spoke with Jiong and found out the problem. Because there are not many frames involved in this flare, I needed to choose a smaller number of them to integrate. Also, the cuts for this flare need to be in a lower range. The highest cut should still include 70% of the footpoints. Although this flare is a little hard to choose a good cut for. I've chosen a range between 5 and 20, as 13 seems to be about average.

    I adjusted the cut range and prime integration based on the first 2 tests, and choose_cut.pro. Jiong reviewed the results after I ran the analysis and said the numbers look very reasonable. So it appears as though I've successfully analyzed one flare from beginning to end (acquiring data, processing images and analyzing the total flux)! Now on to the second one...



    16.june.2006, Fri

    Today I mostly worked on fixing a coalignment problem. I had originally coaligned MDI to an unprocessed image of TRACE. It appeared at first to just be that I didn't shift the MDI image after doing setpoints. After I discovered the true cause of misalignment, I setpoints of MDI to a processed TRACE frame and shifted image1 and image2 correctly. I tested several flickers to make sure the alignments were correct (1 & 2, 1 & various TRACE frames, 2 & TRACE).

    The TRACE data I have processed and saved as one file (040725_trace_final.sav) contains two flares (events 2280 & 2350). So I ran movies to find a quiet period between the two flares and when I process the data I will analyze the two flares separately based on this.

    I also discovered a little bit more about this flare through a quick look at the LASCO CME catalog and another look at the SEC Events data. The flare I have thus far been referring to as 040725, is made up of 2 flares which are event 2280, an M2.2 class flare, and event 2350, an M1.1 flare. Both flares are from the active region 10652.



    15.june.2006, Thur

    This morning Jiong went over the final programs (in "analyse_041107_flare.pro) for measuring flare ribbons and reconnection flux. I ran these programs on the 2004 Nov 7 event (from sunspot 10696, SEC event 4020, X2.0).

    I ran the three programs through and then realized that in order to make good use of the programs I had to find a good cut (which determines the lower limit for the intensity of a pixel chosen to be included as part of the flare). When a good cut is chosen, the contour will be close to matching the boundary of the flare. I used a TRACE frame from the peak of the flare, determined by SEC event data, to run a little program on (choose_cut.pro) to determine a good cut.

    I ran through the programs which map a series of cuts to a coaligned magnetogram (processed earlier) and then integrated the total flux of the footpoints. We can integrate all the magnetic flux encompassed by the footpoints because as the footpoints 'grow' away from each other, the closer parts which appeared earlier are beginning to cool and 'disappear' from our image. We can regard this as reconnection flux because as the magnetic field reconnects...I think I will need to find a picture/cartoon/diagram to be able to describe this. But for now...we're not sure exactly where reconnection is occurring, which is one reason we cannot just measure it directly. I think that another reason it is difficult to observe reconnection is because of its very short time scale.

    Ok, so now with this illustration (disregard the 3 images to the right during the following explanation) I can describe why we can regard integration of all the magnetic flux encompassed by the footpoints because, according to the standard flare model, as reconnection occurs, it causes a heating at the footpoints of only the most recently reconnected field lines (the outermost field lines). This can be seen in the above illustration as the yellow area. As reconnection continues, the outermost reconnection field lines surround the previously reconnected field lines. The previously connected field lines cool, as seen in the red and blue area of the illustration. Because they are cooling, they gradually disappear from the images, which are taken at 1600 angstroms. Therefore, we can integrate the area of the footpoints taken over the life of the flare - the area of the footpoints apparent in the beginning of the flare will have cooled and disappeared from the images by the end of the flare, which will show the footpoints of the field lines which were the last to be involved in the reconnection.

    In the morning I plan to run the analysis on the 25 July 2004 data. After some more experience with the programs, I think that I will be better able to answer the last set of questions.



    14.june.2006, Wed

    This morning I extracted the subframe of the MDI map, rebinned/congrid images to the same (artificle) resolution. Jiong noticed /minutes could not be used for the solar rotation, which fixed the problem I was having of not being able to align the images. Once I coaligned the images, I flickered the two images to examine the coalignment. I found that the MDI image appears to be off by only a couple or a few pixels to the NE of the TRACE image.

    In the early afternoon, I did fine-tune coaligning of MDI and TRACE images by choosing points on each image with setpts.pro. I tried this 3 times and found a satisfactory coalignment on the third try. I saved the smdi_image1 and shifts.

    Next, I extrapolated the MDI image (mdi_image1), which is at the photosphere, to about 2000 km above to the chromosphere where the 1600 Angstrom TRACE image appears. I used a potential magnetic field extrapolation and found that the difference in field strength from the photosphere to the chromosphere in this flare is roughly 500 Gauss - about 40% weaker in the chromosphere! (is this correct?) The images are visually different in that the extrapolated image is 'blurrier.'

    (Note: TRACE = 10 min prior to flare, MDI = 1/2 hour prior to flare)

    (remember to go back and answer questions about height)


    13.june.2006, Tue

    Today, I downloaded MDI .fits files. I adjusted the MDI files for better viewing by resizing them and choosing the magnetic field strength to display. I played with this second option a bit, showing only regions with 2000 Gauss (very few), 200 Gauss (many regions appear) and in between as well. I looked at several frames and noticed that, surprisingly, the active regions shown on MDI magnetograms do not appear to change much at all over several hours. I found a frame that is good to map TRACE data to - it is about 45 minutes before the flare begins.

    I converted my TRACE reference frame (also before the flare) and chosen MDI frame into maps. I plotted these two maps and then overlayed the MDI map on the TRACE map.

    One interesting thing which occurred during this process involves the rotation of the MDI frame. The frame had to be rotated 180 degrees so that the coordinates of the active region being mapped would match the TRACE frame. I discovered that when mapping the MDI image I should not rotate the frame. This is because during the overlay, I was forced to rotate the image. If I had rotated the image when I mapped the MDI frame, the rotation during the overlay would cause the region to be off again by 180 degrees.

    I also overlayed the TRACE image onto the MDI image successfully. However, I was not successful in overlaying the TRACE image to the sub-frame of the MDI image. I will possibly investigate this tomorrow.




    12.june.2006, Mon

    This morning I finished processing TRACE data for 2004 July 25.
    -chose a good reference region and images shifted nicely to same coordinates

    1:19pm - I've requested MDI data starting at least 1 hour before the event. I will choose a time averaged frame from before the flare. This is because during the flare the line profile of the absorption spectrum of the chromospheric region will be redshifted due to accelerated particles into the region, changing the characteristics of the footpoint (temperature and pressure increase).

    In the meantime...

    MDI looks at:
      1. the white light continuum of the photosphere - here we see only the sunspots not the active region
      2. photospheric longitudinal magnetograms - these show us the active regions, measured along the line of sight (longitudinal), which includes diffuse magnetic areas (positive=white, negative=black)
      3. Dopplergrams - these show us how much the chromospheric region has redshifted due to the energy influx from electron beams.

    By looking at the MDI magnetogram we can immediately tell if the active region has sunspots or is diffuse, the location of the active region, the complexity of the active region (alpha, beta, delta, gamma), the area of the active region, and whether the region is being "led" by the north or south "pole".

    Magnetic fields are measured by the Zeeman effect, wherein spectral lines are split (or widened) because the magnetic field breaks electron degeneracy, causing more options of energy levels for electrons or by changing the spin orientation of electrons and thereby increasing their energy. A stronger field means a wider split in the spectrum.

    The equations that are the basis for extrapolating the magnetic field of the MDI (in the lower atmosphere) to the footpoints of the flare in the TRACE data (in the transition region) are:

    F = J x B

    MORE EQUATIONS SOON

    A potential magnetic field is a force-free field wherein j x B = 0 (where j is perpendicular to B, there is no force). Alpha describes how much the magnetic field is twisted, so in a potential magnetic field alpha=0. A constant alpha describes a linear force-free field. Where alpha is neither zero nor constant, there is a non-linear force-free field.

    TO BE CONTINUED...


    9.june.2006, Fri

    Last night I downloaded TRACE data from 25 July 2004 from ~13:30 to 14:00 UT. Today I downloaded a second flare from same date from ~14:19 to 16:30 UT. I hope to process this data over the weekend with a prep_trace program.

    Today I also personalized a prewritten prep_trace program for the 20041107 data as a whole (not as yesterday which was only for 50 frames). I ran this program line by line to understand how each section should work (rebin, shift coord, remove dark current, normalize). Then I ran the program as is. After help from Jiong fixing a few lines in the program, the program ran through to the end.

    I learned a little about shifting for solar rotation after noticing that adjusting for pointing shift did not fix all shifting.

    I also tried to write a little loop program, unsuccessfully. Soon.




    8.june.2004, Thur

    Yesterday and today I practiced with an event which has already been analyzed.
    This included:
      - finding and downloading TRACE data of the event
      - finding a good time frame for data based on SEC Event files and GOES X-ray plots
    Processing included:
      - finding a good (clean) reference frame
      - shifting all frames to match coordinates of good frame
      - resizing (rebin) the images
      - removing CCD noise (dark current) and normalizing the images to the exposure time
    I did this imaging procedure for only the first 50 frames (1 .mxf file). I ran a movie of the images after each process to view the changes.