Virtual impactors: Search and destroy

If for an asteroid which has been observed only over a short are and then l ost there are orbits compatible with the observations resulting in collisio ns, recovery would be desirable to decide if it will actually impact. If re covery is essentially impractical, as is the case for many small asteroids in the 100- to 500-m-diameter range, the next best thing is to make sure th at the lost asteroid is not on a collision course. We propose a method for achieving this guarantee, with an observational effort far smaller than the one required for recovery. The procedure involves the computation of an or bit that is compatible with the available observations and, by hypothesis, results in an impact at some later encounter; this we call a virtual impact or (VI). The collision at some future time is a strong constraint; thus the VI has a well determined orbit. We show that it is possible to compute for each given time of observation the skyprint of the VI, that is the set of astrometric positions compatible with an impact (or a near impact). The sky print needs to be scanned by powerful enough telescopes to perform a negati ve observation; once this has been done for the skyprints of all VIs, colli sions can be excluded even without recovery. We propose to apply this proce dure to the case of the lost asteroid 1998 OX4, for which we have found orb ital solutions with impacts in the years 2014, 2038,2044, and 2046. Suitabl e observing windows are found when the VI would be close to the Earth in 20 01 and in 2003, and the corresponding skyprints are small enough to be cove red with very few frames. This procedure might become more and more necessa ry in the future, as the number of discoveries of small potentially hazardo us asteroids increases; we discuss the general principles and the validatio n procedures that should apply to such a VI removal campaign. (C) 2000 Acad emic Press.