Duck and Cover

[I considered "Don't Look Up," but I have a soft spot for Red Scare propaganda.]

This week there was news of an asteroid with a 1-in-82 chance of hitting the Earth in 2032. Most reports seemed to spin this as a reassuringly low chance, but as someone well-versed in low probabilities, I found it to be uncomfortably high, so I wanted to dig into it a bit. NASA's Small Body Database has an entry for this asteroid, which gives its orbital parameters and a nice interactive plot of its path through space. In order to do some calculations with this info though, I found a tool developed by MIT called poliastro, which can connect to the database and read the orbit definitions.

As a first pass, I tried duplicating the type of plot NASA included on its page, showing the orbit of the asteroid compared with a few of the planets around the time of the possible collision:

This shows the orbits of Earth in blue, Venus in cyan, Mars in red, and the asteroid in magenta. The planets' orbits are all fairly circular, while the asteroid's is sharply elliptical, and crosses the paths of the planets. The predicted flyby is around December 2032, and if you watch the animation you can see the asteroid skirts behind the Earth with a fair gap.

All predictions include errors though, so what happens if this orbit isn't quite the right one? Poliastro offers an object called a maneuver that can perturb a given orbit. Starting with the orbit above, I applied a random change in velocity with magnitude up to 10% of the initial value. Then I measured the distance between the Earth and the asteroid in time for each variation in the orbit:

The x-axis is the Modified Julian Date, a measure of time often used in astronomy. The black line is the predicted orbit, which only gets down to around 200 million km separation. For reference, that's about 1,000 times the distance from Earth to the Moon. However, some of those variations get a lot closer. To see how many of them are getting close, we can look at a histogram comparing the closest approach for each of the variants to the original:

Values below zero indicate the variant gets closer than the base prediction, and the red line marks the distance that would result in a collision with the Earth. Several of those variants are getting awfully close, so much like my pressure cooker post, I'm not sure I feel better at the end of this!