2004 MN4, Even Higher Probability

phreakuencies writes "Worried since the recent post about the MN4 2004 asteroid, I added a bookmark to its 'impact risk' section at NASA. The asteroid started as having a 1/233 probability of hitting earth. Later it raised to 1/63. Daily computations made on 25 Dec raised its chances up to 1/45. Optimists can now say it has a 97.8% probability of missing earth." And Veteran writes " NeoDys offers the 'Orbfit' software package (source code released under the GPL) which can be used to get a pre-release view of the situation with Asteroid 2004MN4."

Comments

Party like it's 2099

[IO ERROR]

And now it's 4 on the Torino Impact Hazard Scale.

The way I see it, we've got about 24 years to party before the world ends. Have another glögg!

Seriously, if it hits 5 or greater on the scale, then we'll have reason to really worry. In the meantime, it's sufficient to just watch and see what happens. As phreakuencies pointed out, right now there's a 97.8% chance of absolutely nothing happening.

Re:Party like it's 2099

[TimToady]

Sorry, it's aimed mostly at the eastern hemisphere. If it hits, it'll be at about 9:22pm in London (.89 of a day at UTC), and since the rock is coming in almost directly from the night side of the planet, it's mostly aimed about 3 zones east of London plus or minus 6 hours. Or were you thinking that Iraq would have been renamed "Texas" by then?

Re:Party like it's 2099

[Scarblac]

Of course the statistics change - more measurements are being done, so the error margin on the estimate goes down. There is a band of possibilities, a bunch of possible trajectories, and more measurements make that band smaller.

Currently Earth is still within that band, and Earth's diameter is about 1/45th of the width of the band, so that's the probability of a hit.

Since more measurements are being done, we'll see this for a few more days - either the band is smaller and Earth is still within it, which raises the probability, or the band is smaller and Earth isn't in it anymore, and the probability drops to 0.

Saying this is "fiddling of statistics" is an insult to the mathematicians involved.

(Story above is simplified, by making it 2D instead of 3D, and by ignoring the fact that it's probably not some fixed area but the probabilities of the thing going outside the area are smaller and it's some weighted average, and I don't really know anything about the maths, I just think it's obvious that the probability would change quickly).

Impact calculator

[Nine Tenths of The W]

Just to reassure you
http://janus.astro.umd.edu/astro/impact/
The impact comes out as somewhere between 450MT and 1.6GT, depending on speed and composition

Re:Something to bear in mind

[barawn]

That's not necessarily true. It depends on the characteristics of the error.

If the errors are Gaussian, if the nominal trajectory (i.e. "it misses the Earth by X+/-Y km") is accurate, but imprecise (that is, X is correct, but Y is large compared to X) then the probability of impact will decrease as the precision is improved (i.e. as Y decreases) because the "Earth impact" possibility moves farther out on the fringes of the observation, and the area doesn't shrink fast enough to compensate for this.

Of course, if the errors are flat (all solutions are equally likely - actually, if the PSF falls off slower than the area shrinks) then you're correct. I'm pretty sure that they're Gaussian, or approximately Gaussian, though. So the only way the probability could be increasing is if the nominal trajectory's impact parameter is decreasing - that is, closer impact.

Re:Common mistake in press coverage

[angel'o'sphere]

Well, seems you don't get the problem.

Simple map of the Earth moon Sysstem:

1 _ _ _ 10 _ _ _ 20 _ _ _ 30 _ _ _ 40 _ _ _
E ... m

Simple map of the path of the asteroid:

/ / / ... ... / / /

That should be 42 stripes but the junk char filter ....

As you see, the first stripe hits earth, the other 41 don't hit. Well, moon is in fact at opposite position when the asteroid comes in, but it was difficult to "draw" that. So, remember moon wont be hit.

Further observations of the asteroid will give us more data to determine wether the asteroid will travel stripe number 1 or 2, or wether it will travel stripe number 20 or 30 or whatever.

If we figure the asteroid is traveling NOT stripe 1 we are 100% certain that it will miss us.

If we figure the asteroid will not travel stripes 31 to 42, the likelyhood of an impact increased to 1:30.

Both calculations are "100%" certain. OTOH, your parent was right. The likelyhood that the chance of getting hit decreases is high. You have 42 draws ... and in 41 draws you have the chance that stripe number 1 -- the stripe which hits earth -- is removed from the set or possible pathes. Because more accurate measurement shows that the asteroid wont go that path/stripe.

angel'o'sphere