[ExI] (NASA.gov) NASA to chronicle close Earth flyby of asteroid (fwd)

[Anders Sandberg]

On 16/02/2013 16:32, spike wrote:
> OK, assume it into existence.  For the sake of thought experiments, assume
> any advance warning you wish, and any existing or feasible future rocket you
> want.

Any advance warning? Now you are making things too easy, Spike! You just 
need to nudge the asteroid away from a collision keyhole, and that can 
be done with a mere kick if you have enough time. The displacement goes 
up linearly with time, except for near Earth passes where it can be 
amplified a lot. So have your rocket just hit the asteroid straight.

(Keyholes are fun. They are the pre-images of Earth along a transversal 
manifold: go through one, and your orbit will hit it next time. Besides 
the 1-orbit keyholes there are pre-images of pre-images, producing 
smaller and distorted keyholes for more remote collisions. They form a 
Cantor-set like fractal along a curve in the manifold. Aim your kick so 
that you miss the fractal - possibly making use of the infinite subset 
of repelling periodic points embedded in its border - and you are safe. )

Looking at the graph on
http://www.neoshield.net/en/mitigation-measures/overview.htm
suggests that if you only have a few days, then either you must already 
have nuclear defenses in place in space, or you need to do duck and 
cover civil defense. Figure 2.2 in 
http://isulibrary.isunet.edu/opac/doc_num.php?explnum_id=142 has more 
detail, and doesn't assume an impressive space shield (I love the option 
"continental relocation")

>    Now what?  If you had in mind putting a nuke aboard and trying to
> blast the rock into gravel, keep in mind the relative velocity.  We have no
> current technology that would detonate a nuke with a tight enough precision
> to do much to a meteoroid.  For instance, if we manage to get the trigger
> tech to plus or minus one millisecond, that is plus or minus a couple of
> kilometers.  If the nuke is more than a few meters away at detonation, the
> rock would scarcely notice, and even if we manage a perfect detonation, it
> isn't clear to me it would break up the meteoroid.
Yes, nukes in space are surprisingly wimpy. But I think you are too 
pessimistic about detonation precision: typical space velocities are on 
the order of several kilometers per second. So a millisecond error is a 
few millikilometers - a few meters. You don't want the detonation to be 
on the surface if you aim for heavy deflection: the blast causes surface 
ablation that does the actual pushing. Groundbursts apparently cause a 
jet of faster but lighter stuff, imparting less momentum.

The big problem is that nobody knows how asteroids do plastic 
deformation in the case of trying to nuclearly disrupt them: gravel 
piles can just mould themselves into a new shape. And they can easily 
just turn into a vast cloud of impactors that will now damage a large 
part of Earth instead of a small corner. Some simulations I have seen 
produced very impressive messes.


-- 
Anders Sandberg,
Future of Humanity Institute
Oxford Martin School
Faculty of Philosophy
Oxford University