Saturday, January 29, 2011

Maybe a planet

kw: science, citizen science, astronomy

In the past few weeks, I've learned to classify stars released by the Kepler Mission to the Planet Hunters project in Zooniverse. Two sayings of Thomas Edison come to mind as I check the light curve of star after star for the telltale feature that signals "planet here": "Invention is one percent inspiration and ninety-nine percent perspiration" and "I learned hundreds of things that don't make a light bulb." To date I've checked about 2,200 stars.

The Kepler instrument records the brightness of about 150,000 stars every thirty minutes. At the Planet Hunters web site, a thirty-day supply of such data are kept and parceled out, star by star, to interested hunters such as myself. The vast majority of the stars have a "curve", really a collection of dots representing the brightness measurements, that looks like this one, except for the little dip at the right end.

This is a typical quiet star, though there is a flare at day 4. The scatter in the data arise from the statistics of photon counting when you are taking short measurements from a 14th magnitude object. As you can see, nearly all the data are confined to the range 1.0075-1.0085. The band of light that draws the eye is mostly in 1.0078-1.0082. This star is a K type star of the same radius as our Sun, so it is probably K0 or K1. A planet the size of Earth would intercept only 0.0001 of its light (0.01%). Now look at the dip near day 30 (on the X axis). Its depth is about 0.00025, so if this is truly the transit of a planet, its size is 1.6 times that of earth. It is even more exciting that the transit takes so long. See the next image:

By counting the dots, I find 34, which means the transit took 17 hours. A transit of the Earth across the Sun, as observed from a nearby star, would take 13 hours, so the velocity is about 75% of Earth's about the Sun. This is a lighter star, so the orbital radius will be similar, but this is also a dimmer star, so I'd say it is near the outer edge of the star's habitable zone, perhaps in a 500-day orbit, comparable to a spot halfway from Earth to Mars in our solar system.

That is a lot to infer from 34 dots on a graph. It may be that there is no planet and this was something else. But it is exciting to realize that some of the features like this one, as seen by "citizen scientists" and cross checked by the project scientists, will indeed signal planets about some of the stars. They eventually expect to find hundreds.

I did a few rough calculations about this project. The chances of finding any random planet about a star of type F, G, or K, in an orbit ten days or longer (shorter ones don't interest me much) is one in thirty. If every star has planets, then the project as a whole could find five thousand. However, the chances of seeing a planet in the habitable zone are quite a bit smaller. For a G star, the chance is about one in 600. Double that (1/300) for a planet of a K star, about half of that (1/1,200) for a planet of an F star. K stars dominate, so overall, the number of planets at a habitable distance from their star, that we can detect with this project, is likely to be about 400 or 500. This number will be smaller if any substantial number of stars are totally planet-free. I count that unlikely.