Thursday, June 17, 2010

Cultivating Your New Experience

I've been bad at posting daily, but I will get better. Since Monday, update, I've read the RR Lyrae and BHB star papers. The RR Lyrae one had V-band magnitudes, which was useful as a jumping-off point, even though it's not one of the LSST filters. The BHB paper has magnitudes as a function of metallicity, which gave me some pause because the metallicities in the simulation are higherthan expected. For now I'm using RR Lyrae stars as tracers.

I also had some difficulty resolving the proper-motion discrepancy, which turns out to have mainly been caused by my own negligence of units. Now that everything's in MAS per year, the differences are much smaller (though not quite zero). I'm both happy and frustrated about that, as I spent a good deal of time trying to find some deep underlying cause when most of the problem was simple oversight. Now that the proper motions are in better shape I've started to convolve them with LSST errors and make plots of "observations." I'm not 100% on how to make plots of vx versus vy for different radial bins other than by oplotting each bin onto the graph, so that's teh next thing to learn about IDL.

As far as the FoS simulation is going, it seems like starfish and x11 don't get along so well, and Joe Cammisa has been great in helping out with that so far. I've also started to try binning the halo particles into 'pixels,' which ran into a snag with the where() function, which now believes that all of the stars are in all of the bins. I also looked at the code for smoothing filters and am a bit unsure of it. The FoS idea seems more intimidating the more I think about it.

Monday, June 14, 2010

Yes yes, yes. Without the oops...

Today was semiproductive. I started off with a simple test to see if the proper motions I'd calculated for the star particles was right: I added the components together (vector-wise), that would give the magnitude of the total proper motion. Another easy way to get this is to take the total velocity and subtract off the radial velocity (again, vector-wise). These should give the same numbers, but unfortunately do not. I spent most of the morning finding little problems with my definitions of each component, but this didn't seem to help. In the afternoon I emailed Adi and Beth for help, but so far there's still a (fairly large) discrepancy.

I also read through the manual for StarFISH (available here: and am going to start playing with it and seeing what's what.

On the astronomy side of things, I looked through ADS for some papers on absolute magnitudes of BHB, RR Lyrae and RGB stars. I found a couple that could be valuable and am reading through them now. I'm glad I have something to balance the computational side of things right now, because otherwise I'd go mad from all of the debugging. But now that I'm getting absolute magnutudes for tracer stars, I can start plugging them in with the estimated errors for the LSST and make some proper motion plots (once I get that straightened out).

Thursday, June 10, 2010


I've been neglecting my blogging duties, but am now back and better than ever. I've got my velocities de-composed into radial and proper motion flavors, a nice aitoff plot of the galaxy (at least, its disk and bulge), have read about StarFISH and the Field of Streams, and am staring to investigate isochrones to generate stellar populations in the simulations. I've also found the interparticle spacing for the halo at different distances from the galactic center, and have a measure of the physical pixel size in the FoS image, which is, thankfully, smaller than the interparticle spacing. The next steps are to get some proficiency with StarFISH so that we can see if it would be a good way to generate stars for the simulation, and to start doing kinematic "observations"of the simulation.

The kinematic studies are on a slightly different track than I've been on. Recently I've been thinking mostly about a FoS replica and whether that would be feasible, but this will definitely be doable, and should be less complicated. Short-term analysis is looking at proper motions of the particles at different radial distances, searching for differences between accreted and in-situ stars. I'll then do that again incorporating the LSST measurement uncertainties into the 'observations.' Before doing that, however, I'll need to think about stellar tracers (BHB, RR Lyrae and RGB stars) and how to model them in the simulation without generating cmds. More on this as it develops

Friday, June 4, 2010

The Arctic Monkeys

It's been a fairly productive past few days. I've got stars in Radial Distance, RA and Dec instead of X, Y, and Z (as well as galactic latitude and longitude), decomposed velocities into proper motions and radial velocities, read about the "Field of Streams" and halo formation models, and I learned (thanks to Maya) how to write out a fits file with all of the newly calculated data. At least, I think I did that last part correctly. Tonight I'll be testing that, but I'm confident that it worked.

Wednesday, June 2, 2010

The Decemberists

So, today was fairly productive. I skimmed a few papers about RR Lyrae frequency in globular clusters, and quickly determined that an S of 5.6 meant an average of 5.6 stars per cluster luminosity, so now those numbers make sense. I also saw some of Stephon's talk, but had to duck out after an hour when I figured how to fix the coordinate change I'll go into below.

The rest of the day was spent working out a coordinate transformation to change the perspective of the simulation to one inside the disk of the galaxy. Basically, I'm re-centering the galaxy and then converting from x, y and z to R, Right Ascension and Dec. I think I've got that solved (though I worry it may be too simplistic), but I won't know until I can fix up a rogue for-loop, which shouldn't take too long, but I'd like to eat dinner and see some sun first. Next on the plate is decomposing the motion into radial velocity and proper motion, and then rotating the galaxy so when we do change perspectives everything's aligned. In addition, I've got some nice reading to do about the "Field of Streams" among other things.

Tuesday, June 1, 2010


Today was a bit of a disappointment. I spent the morning going through the Catelan paper on Horizontal Branch stars looking for something about specific frequencies, but all I could find were a few metallicity-morphology relations. After looking through some of the references, I decided to just to an abstract search on ADS for "specific frequency of horizontal branch stars" and "specific frequency of RR Lyrae stars." From the latter search I found a promising paper: "RR Lyrae Stars in the Andromeda Halo from Deep Imaging with the Advanced Camera for Surveys," which has some good information on metallicity and classification that helped my understanding. Unfortunately, it contextualizes its results by pointing out the debate and differing results on this subject, so it's not a settled issue. Also, I'm not sure what definition of specific frequency is being used in the paper, as the number they give is 5.6, which is not what I was expecting. So, that bears some investigation.

Also, I made a nice plot of uncertainty in proper motion (in km/s) versus distance(in pc) from some listed LSST uncertainties (in mas/year). This should be a simple unit conversion problem, but for some reason both axes run from large negative numbers to 0 instead of from 0 up. It seems like the problem occurs where I initialize my vector of distances, because the only thing I to to them is multiply by positive factors but they turn out negative. That'll give me something to futz with tonight, at least.

*update, it turns out I was doing scientific notation wrong in IDL, which was confusing it. I'd also forgotten to convert from mas to radians, which explained why the errors I was getting were so high. Now things look somewhat more reasonable.