| |
|
Stars orbiting the Supermassive Black Hole at our Galactic Center
For lots of further information you could have a look at this UCLA Faculty Research Lecture by Andrea Ghez (see the chunk between the 7th and 51st minute).
One of the major achievements of the paper was in creating techniques precise enough to follow the motion of individual stars near the center of our galaxy, where the density of stars is incredibly high. Based on these observations, Ghez et al were able to confirm the existence and type of a large black hole there. DownloadMoviesThese have been up since May 2005. The modifications made in March 2006 were just modifications of the captions. Here are 800 x 600 movies in MPEG format. (Updated March 14 2006)
Interactive 4d ModelHere's the Partiview model, it should work on Windows, Linux and OS X. These include GeoWall versions for Windows and Linux. Just unzip the file and read the documentation (or just click on stars.bat if you're running Windows). The stars and their orbits are shown in a box of length ~0.8 parsecs, or about 15 trillion miles. Specifically, the box length is 2 units, where each unit is an "arcsecond distance from the black hole. The center of our galaxy is about 8000 parsecs (or 8 kiloparsecs) from Earth. And at that distance, 1 arcsecond is ~0.04 parsecs = 8000 astronomical units (AU)." (Quote from an email of Jessica Lu, 8 Sep 2004.) You can turn orbits, stars, and a date ticker on and off. Detailed instructions are in the README.html file provided. How this was madeThe researchers (Ghez, Lu) provided the raw data file stars13.dat to Randy Landsberg, Mark SubbaRao and Dinoj Surendran of Cosmus. The file has the x-y-z positions of 13 stars, with every 14th line having the year and time index number. The whole data file has about 15 000 timesteps. This was converted to Partiview format by the not-very-general perl script dat2pv.pl we wrote. perl dat2pv.pl stars13.dat somename K Typing the above produces files somename.cf, somename_stars.speck, somename_orbits.cmap and somename_orbits.speck using 1/K of all timesteps. The speed is changed so the total period of the simulation is the same - in this case, 15 seconds. Improvements are certainly possible; orbits could be made more intelligently so they are smoother and smaller. |