Transiting Planets
Hi I’m Matt, a graduate student at Yale University and a member of the Science Team. We’re really impressed with the turnout so far on planethunters.org and users have already pointed out some really amazing objects! Quite a few people have asked for some clarification on what transits look like, so I’ll address that in this post.
In the figure above, we’ve taken a Kepler light curve from a star that’s about the same size as the Sun and have simulated what the effects would be if a few different types of planets were to transit.
The white dots show the amount of light from the star measured with Kepler with no planets transiting. The blue points show what we would see if a planet just like Jupiter orbiting this star were to transit. This Jupiter-size planet, at about 11.2 times the size of the Earth and one tenth the size of the star, is shown to scale transiting its parent star in the top left blue box.
The green dots show what a planet just like Neptune would look like transiting. Since it is much further away from the star than Jupiter, it would have a slower orbital speed so it would take longer to transit the disk of its parent star, which is what explains the longer duration, or wider width, of the transit event. With Neptune’s much smaller size than Jupiter, at 3.9 times the radius of the Earth, it doesn’t block out as much light, which is why the depth is much shallower.
Both of these events are very noticeable, compared to the effects of an Earth-size planet. The tiny speck on the star in the far right red box shows, to scale, what a transiting Earth-size planet would look like if we could see it. Now you get an idea of how difficult finding Earth-size planets is going to be! If that transiting planet had an orbital period of 1 year just like the Earth, then the dip in light observed from the parent star as the planet transits would be similar to the red points in the light curve. Since the Earth is much closer to the star, it has a much faster orbital speed, which then makes the duration of transit much shorter than the duration of either Jupiter or Neptune. Because the Earth-size planet is much smaller than either Jupiter or Neptune, it also blocks out less light making the dip in light we receive here on Earth barely discernible from no transit at all.
We don’t expect people to see these events all the time, so don’t worry about missing them. That’s why we’ve introduced fake planets into the mix. The fake, or synthetic, planets will help us determine the completeness of Planet Hunters, or how likely we are to detect planets of different sizes and with different orbital periods if they exist.
20 responses to “Transiting Planets”
Trackbacks / Pingbacks
- - December 24, 2010
- - December 26, 2010
- - November 4, 2014
Planet Hunters 
Thanks for the article, that helps quite a bit! Could we have an article (or series?) about eclipsing binaries and variable stars too? They seem to keep cropping up the data!
So far I’ve looked at just over 400 stars but I’ve not seen any with dips that have a flat bottom like the Jupiter or Neptune traces yet. All the possible transits I’ve identified have gone dips that have gone straight down and straight back up again. If they’re real then I guess that would indicate that they’re jovian-sized objects in very tight orbits (so they block out a lot of light, but don’t do it for very long)?
Thanks for the explanation although this already was clear. In the graphics the earth-like planet is hardly noticeable. Are there more real examples of earth like candidates available for study? In short are there more real life examples available of any real transit. In the data we receive to check the vertical scale is “fixed” so for me sometimes it is really hard to make any decent analysis. In your example there is a nice dotted line on the 1.000 reference but in the data received for checking the dots are all over the screen. I just consider them as stars with no transits but would like to know if I am analysing this the correct way.
Thank you Mat,
a very helpful post for me.
Any chance of making this available to be seen/viewed/zoomed etc with the regular tool we’re using to classify the stars. It seems like it would helpful to see what these look like at different zoom levels etc.
(Particularly for close-in planets like Earth, Mercury, etc. I don’t think anyone will miss a Jupiter or Neptune!)
Even if you can’t put it on the regular graphing tool, even just showing some zooms at a few different levels would be helpful.
Just wondering what the transit depth would look like for an earth-sized planet in an orbit similar to earth (assuming non-eccentric). From the chart in your blog it looks like less than 1/10th a percent for a star the size/ brightness of the sun or is it much smaller. Thanks!
Hi, Mark. That’s very helpful.
Both Jupiter and Neptune create a dip that lasts much longer than anything I’ve seen here (which must include plenty of simulated data). Given reasonable assumptions, what’s the shortest dip we should consider as a potential transit? Clearly one low point does not a dip make. But two? three?
” . . . what’s the shortest dip we should consider as a potential transit? Clearly one low point does not a dip make. But two? three?”
As you can see from the graph above, the length of the transit is directly related to the distance from the star and thus to transit frequency.
So certainly a 5, 4, or 3 point dip (or even a 2 or 1 point dip) COULD be a transit. But if it is, you are certain to see the transit repeat just a few hours or days later, and continue repeating at a regular interval thereafter.
So for the short-duration dips, you are looking for a combination of the dip plus repeatability.
I’ve been working to figure out the period of some examples, so here are a few:
APH10106224 – period 4.16 days, 8 transits visible, length of transit 3-7 dots (about 1.5-3.5 hours)
APH10050255 – period 1.285 days, 25 transits visible, length of transit 2-4 dots (about 1-2 hours)
APH10050255 – period 0.643 days, 51 transits visible, length of transit 1-3 dots (about 0.5-1.5 hours)
APH10115270 – period 0.537 days, 62 transits visible, length of transit 1-6 dots (about 0.5-3 hours)
APH10086257 – period 3.23 days, 10 transits during the quarter (though only 5 were definitely visible, 2 were just 1 point, and three were missing in action, so this object is a bit more dubious than the others), length of transit 1-4 dots (about 0.5-2 hours)
To summarize: As a practical matter what you are looking for is short transits that repeat within a few days. If they don’t repeat, you can ignore them.
Longer transits won’t repeat within one quarter, so all you can do is mark it and wait for further data to confirm.
Personally I’ve been marking all longer transit candidates, but the short ones only if I can find a regular repetition.
Also as a practical note–if I see a couple of 2 or 3 point dips, I start measuring the distance between just using my fingers on the screen (you could, of course, use a ruler or something similar). If I can find 3-5 potential transits in a row that appear to be the same distance apart using that crude method, they have pretty much always panned out on further investigation.
Obviously you have to try out dozens, or maybe into the hundreds, of candidates before finding any keepers–but on the other hand I ran into a couple just tonight, just in the ones presented to me as I’ve been classifying.
So these short dip/high frequency transits are out there if you keep looking for them.
This is helpful, but I wonder why the examples are so thick in the y scale and broad in the x, compared to what I have been seeing and marking as possible transits. What I have been marking are drip like dots dropping from the main light curve. I haven’t seen anything that looks like the examples in this posting
This link at nasa shows what transits in our own solar system would look like:
http://planetquest.jpl.nasa.gov/transit/indexTransit.html
@Brent
If you use the equations given at http://kepler.nasa.gov/Science/about/characteristicsOfTransits/
you can derive the approx. transit duration using a formula like: Td = 13 x (P^2 x Ms)^0.167
which works out to around 3 hrs so for APH10106224 you’re in the right ballpark for a transit
This is brilliant! It should definitely go in the tutorial.
Goodness. Earth is tiny. Our detection techniques better step up their game.
קידום דפי מקוונים עושים רק בחברת קידום אתרים מקצועית כמו חברת נט-סטייל . צרו עמנו קשר על מנת לרכוש שרות של קידום אתרים בגוגל בדרך הטובה ביותר.
this is useful.
we know that our sun goes in solar cycle in 12 years during which the intensity of light varies which is periodic.so it can also be the case with other planets and so how can we distinguish these from transits
Thanks Matt,
It’s useful to know what a ‘real’ system would look like.
I’ve been flagging a number of single point dips if they are significant enough. From this, I would say they are noise, though. Oh well.
@raj: A 12 year cycle of variability being shown on a quarterly basis would look pretty much as flat as it’s shown here.
The real problem is that a star as quiet as our own still has significant variability when you scale the Y-axis, and it would be pretty easy to lose the transit of earth in that variability. But Kepler would still detect that dip.
In the graph here, it *seems* that the amount that an earth-sized planet at 1 AU would make the measurements dip is about 0.0005 magnitude, lasting about a full day. But that’s really hard to see in this graph, since we can’t zoom in on it like we can our normal data. One thing for sure is that single point dips sure aren’t earth-like. 🙂