Saturday, November 30, 2013

"What was that bright light in the sky last night?"

While looking for information on comet ISON, I ran across an interesting project by NASA: the All Sky Fireball Network, a network of (currently) 12 black-and-white video cameras able to image the whole night sky.  The images from these cameras are processed by ASGARD (All Sky and Guided Automatic Realtime Detection), developed for a similar effort in Canada.  These astro folks sure love their acronyms.

A "fireball" is any meteor brighter than venus.  These are not rare events.  There are several on most nights, and nights with a dozen or two are not uncommon.   As there is deliberately quite a bit of overlap in the cameras' fields of view, most fireballs are caught on multiple cameras, making it possible to calculate the three-dimensional trajectory of the meteor, and from that determine its orbit. posts a graphic of the night's calculated orbits.  I believe they produce that themselves from the raw data on the NASA site, which just seems to show the orbital elements as text.

It's pretty easy to see why this data is of interest to NASA.  Not only does it provide information about the composition of the Solar System, it's of practical use in designing satellites.  But even without that, it's a just plain neat hack.


earl said...

v. cool. And totally head-swimmy. The model of the solar system we were taught in grade school (ok, I went to a very small grade school) was basically the ball and string model, and we thought of orbits as circles. Later, we learned that they were really ellipses, and later still that in an idealized two-body system both bodies were in elliptical orbits around (one focus of which was? See, I'm already confused) their mutual center of gravity. But with hideously eccentric orbits, like that of ISON and some of the green orbits in the hack you cite, both objects interact for millennia with other objects and may even change their mass as well as their position. It would seem that predicting the return of a comet a hundred years or so hence must require about as much faith as math.

David Hull said...

On the one hand, yes, orbits can get very complicated. There are various orbital resonances to deal with, for example Pluto's 2:3 resonance with Neptune. There are asteroids (e.g., 2006 RH120) that switch back and forth between orbiting the Earth and orbiting the Sun directly. There is good reason to believe that Neptune started out in a much lower orbit and moved out to its current position by (I think) exchanging angular momentum with infalling debris. See the Nice (as in France) model for further details.

On the other hand, as I understand it, you can get pretty far by looking at the influence of the Sun and probably Jupiter, then doing a finer calculation if you find your object is going to pass near one of the other planets. Even some of the funky orbits I mentioned above are reasonably predictable. Resonances form precisely because they're stable.

When it comes to predicting the orbits of comets and other things that could hit the Earth, I believe we have a decent handle on it. I've read descriptions like "Asteroid XYZ will pass at twice the Earth-Moon distance on this orbit, +/- X%. There is a Y% chance (usually really small) that it would hit Earth on its next orbit in Z years." In such cases we know a lot more after the passage, and Y% generally gets even smaller.

As I understand it, there are often "keyholes", which, if the object hits them on this orbit, they're much more likely to be at a certain point on the next. Without looking at the details, I believe these are places where some particular energy function they're looking at has a high second derivative, meaning that a small change in position around there can mean a large effect down the road, so you know much more after an object passes through such an area than you know before. Out of such non-linearity comes chaos and, conversely, most of the time things are fairly linear and predictable.

Most of the time.