This week I was introduced to the program by all the professors and met my advisor, Andrés, along with Dana Longcope to go over the general scope of the project. Alongside this every morning we met with John Unverferth to learn the IDL programming language which helps us to work with much of the data. First impressions of IDL: don’t like it. The syntax seems inconsistent and it is not very user friendly. Note I am coming from Mathematica, a very concrete language and is also extremely interactive in terms of the data.
After Keiji helped us get our computers set up with the school’s network, Andrés showed me the program we were going to be using for the summer. It uses the Bipolar Active Region Detection (BARD) code, and explained to me how it works. Loading up a magnetogram, the program searches for cores or kernels above a certain user-set threshold and circles them. It will erode them away to eliminate any stray cores and then regrow them. Merging is also included in the process to make for a more consistent magnetogram. The program will also search for these kernels in layers, decrementing the kernel threshold each time. All of the parameters are customizable; thus the first step in our process was to optimize the parameters for the data I was going to be working with. The previous REU student had optimized the code for use with the MDI magnetograms. I was working with the magnetograms from the Kitt Peak Vacuum Telescope (KPVT). During this first week Andrés helped me set up a code to run a combination of kernel thresholds and erosion and dilation parameters. By looking at the pictures and determining which one did the best job. You can see from the pictures below that the one on the left has a very low kernel threshold (200 Gauss) and circles a lot of regions. This is cumbersome and undesirable. The picture on the right has a high kernel threshold (600 G) and does not circle enough. The goal: to find a happy medium between all the parameters.
On top of this, a lecture was given by Andrés on the solar cycle and dynamo.
|