Slashdot Mirror
NASA Downgrades Asteroid-Earth Collision Risk
coondoggie writes "NASA scientists have recalculated the path of a large asteroid known as Apophis and now say it has only a very slim chance of banging into Earth.. The Apophis asteroid is approximately the size of two-and-a-half football fields, and updated computational techniques and newly available data indicate the probability of an Earth encounter on April 13, 2036 for Apophis has dropped from one-in-45,000 to about four-in-a million, NASA stated."
According to NASA, it is 1 in 135,000 and diameter is 0.270 km
http://neo.jpl.nasa.gov/risk/a99942.html
The image on TFA gives the impression that it is way larger than 20km and the summary claims that is is 200 yard = 0.182 km. And the text claims that it is four-in-a million aka 1 in 250.000.
Posted too soon, I did. Should have checked the source of the picture first.
What we have accompanying the article on Apophis is a picture of Asteroid 253 Mathilde. Apparently pulled off the NASA website at random by the author's of TFA.
"I do not agree with what you say, but I will defend to the death your right to say it"
All the models are run according the 'standard dynamic model' at JPL which includes gravity from the Sun, planets, large moons and large asteroids. Perturbations caused by objects outside the solar system are negligible compared to non-gravitational effects such as direct solar pressure and the Yarkovsky effect. These effects are impossible to model without knowledge of material composition, mass and structure, which you can't really get without going there.
However, the uncertainty caused by these non-gravitational effects is very small compared with the uncertainty caused by the fact that we just plain don't know quite where it is and how fast its going. In order to know where the asteroid will be in 2036 to within an Earth radius requires us to know where it is now to within about a meter -- the 2029 close approach in particular magnifies uncertainties incredibly (100x).
These state estimate uncertainties overwhelm any small errors in the dynamic model, and these new and improved probabilities come from refining the current state estimate. So yes, it is still valid to make these kind of predictions. You have to start early (10-20 years) to be able to stop it as well, so its important to keep an early eye on it.