1950 DA could hit Earth in 2880, but do we really know it? Yarkovsky effect could make a big difference in calculating its whereabout in the next centuries. So we maybe better think of....

The sweet solution
by Andrea Milani - Copyright Tumbling Stone 2001
From our correspondent at the Asteroids III meeting, Santa Flavia (Palermo)


Many of the presentations in the second day of this meeting (see T.S. presentation of the second day's talks and the related topics) were about a bizarre phenomenon, which goes under the name of Yarkovsky effect (dict.). A Russian engineer first proposed this idea in an article published in the early 1900s, but this paper has now been lost. The suggestion by Yarkovsky was revived by Opik, an Estonian astronomer working in Northern Ireland, in the 1940s.

The Yarkovsky effect can be described in this way. An asteroid orbits around the sun with a period, the asteroid year, typically longer than an Earth's year. It also spins with a period, the asteroid day, tipically shorter than our day. Thus the asteroid has seasons, with a "summer" in which the northern hemisphere is more illuminated, thus hotter; it has also "afternoons" warmer than "mornings". The warmer half of the asteroid emits more thermal radiation; since radiation carries away some momentum, this propels the asteroid like a rocket. Of course this radiation effect results in a very gentle push, and it changes the orbit of the asteroid very slowly. However, if the same minute push continues to act for millions of years, the asteroid can move a long way from its original orbit.

According to a theory pioneered by the late Italian astronomer Paolo Farinella, this can result in transporting small asteroids from the main belt (dict.) to a Near Earth orbit. That is, radiation effects like Yarkovsky can result, in combination with other known dynamical mechanisms, in the delivery of meteorites to the Earth. This subtle effect is the basis for several of the studies, presented today, on the long term evolution of the asteroid main belt and on the origin of meteorites and Near Earth Asteroids (dict.).

An even more remarkable implication of the Yarkovsy effect has been presented today in a poster paper by Jon Giorgini and many coauthors (from both the Solar System Dynamics group and the asteroid radar group of the Jet Propulsion Laboratory, California). The paper is on the asteroid 1950 DA, discovered by Wirtanen during a close approach to the Earth in 1950 and never seen since, until it was recovered by the Lowell Observatory Near Earth Objects Survey in December 2000. In March 2001 this asteroid was observed with radar (dict.) from Arecibo and Goldstone (see T.S. special issue) by Ostro, Hudson, Benner, Nolan, Margot and Giorgini, allowing them to reconstruct the shape and the rotation, and also to compute an extremely accurate orbit.

As a consequence of this extremely accurate orbit determination, Giorgini, Chodas, Chamberlin, Chesley and Yeomans were able to compute the orbit of 1950 DA for a long time span in the future, until the year 3000 AD. During this computation they found that in the year 2880, more exactly on 16 March of that year, a collision with the Earth could not be excluded.

Since 1950 DA is more than one kilometer in diameter, this was a serious case to be investigated, although some may argue that what could happen in the 29th century is not our immediate concern. It is, however, a very important test of our present capabilities of prediction. Thus Giorgini et al. have studied the following important issue. Is the dynamical model we use to propagate the orbits of the asteroids accurate enough to allow a definite prediction for an event in 2880, with an accuracy better than the diameter of the Earth? The answer to this question had to be negative, and the main reason is the Yarkovsky effect. The orbital changes resulting from the Yarkovsky>effect have never been directly measured, but a model for their orbital effect is available and can be used to estimate its magnitude. Actually this model has been developed by a former coworker of Farinella, David Vokrouhlicky (Charles University of Prague), in collaboration with Steve Chesley (while he was still at the University of Pisa, before moving to JPL) and myself. By using this model, Giorgini and Chesley found that the Yarkovsky effect can change the position of the asteroid 1950 DA at the critical date, 16 March 2880, by as much as 20 million kilometers. This also implies that the probability of an actual impact in 2880 is much less than the probability of being hit before then by an as yet undiscovered asteroid.

From this study it is possible to draw two rather startling conclusions. One, if we were interested in predicting long in advance, e.g., now, the outcome of the 2880 encounter of 1950 DA with the Earth, and decide if it can actually hit, we could not do it. To be able to achieve such a capability of prediction we need to improve our understanding of the Yarkovsky effect: this implies that we must be able to model quite accurately the temperature on the asteroid, and how it changes with the local hour and season!

Second, if indeed some day, maybe in the next century, mankind were able to establish that the asteroid, left to its natural orbit, would impact in 2880, then a surprisingly simple solution would be available to deflect it. To change the size of the Yarkovsky effect we need to change the temperature of the surface of the asteroid. This can be achieved by changing its color, in such a way that more of the solar radiation is reflected: that is, the asteroid needs to be painted white. Since the Yarkovsky effect is much larger than the diameter of the Earth, a small change in color would be enough, such as painting only one thousandth of the asteroid surface. This could be achieved, rather than by paint, by spreading on the asteroid surface some very fine white dust. A back of the envelope computation shows that only about one ton of white dust, such as powdery sugar, could be enough! This would be a comparatively cheap and sweet solution.