Source - Space Daily
by John Rehling Los Angeles CA (SPX) Oct 14, 2013
by John Rehling Los Angeles CA (SPX) Oct 14, 2013
 The algorithm is overly lenient in interpreting normal variations in starlight as being transits of small planets, and this was further complicated by abnormalities in the Kepler instrument itself. To explain this, it's necessary to say something of how Kepler operated. Figure 1 and Figure 2 larger size. (file image) An artistic view of the system seen from Kepler-62f. The host star is slightly redder than our sun. The smaller exoplanets Kepler-62b (1.3 times Earth's radius) and Kepler-62c (0.5 times Earth's radius) are close to the star. Kepler-62d (2 times Earth's radius) is significantly bigger and closer, Kepler-62e (1.6 times Earth's radius) and Kepler-62f (1.4 times Earth's radius) are relatively close to each other and both are sustaining water and rocky surface as suggested by the clouds' color, water, atmosphere and rocks Credit: Danielle Futselaar/SETI Institute. For a larger version of this image please go here. |
A new analysis of observations from the Kepler spacecraft reveals what may be the first earth-sized planets with earthlike temperatures found orbiting sunlike stars.
Until now, Kepler's nearly continuous observations of over 150,000 stars have confirmed the existence of Earth-sized planets in the hot regions close to their star. Larger planets, some as small as one and a half times the Earth's diameter, have been found in the Habitable Zone, where the amount of heat they receive from their star may sustain earthlike temperatures.
In addition, earthlike planets have been found in the Habitable Zone of tiny, cool red dwarf stars, which may offer a more hazardous environment for life than with sunlike stars. But finding planets with the combination of earthlike size and earthlike temperature around sunlike stars, a major goal of the Kepler mission, has been elusive.
Part of the challenge is how difficult it is to spot a single transit by an Earth-sized planet, which blocks only 0.01% of its star's light, far less than larger planets block. This figure (figure 1 top image) shows the graph of three different stars' brightness as measured by Kepler during actual transits by extrasolar planets the size of Jupiter, Neptune, and Earth. Transits of larger planets stand out prominently, but transits by Earth-sized planets are so subtle, Kepler can only identify them when there is a series of many such dimming events occurring at the same regular interval.
Last December, the Kepler team released a list of 18,406 possible planets found by an algorithm that searched the first three years of Kepler data looking for series of transits.
If we define a size limit of no more than 1.25 times the Earth's diameter and a Habitable Zone where the amount of heating that a planet would receive is between that of Venus and Mars in our solar system, then this list contains 87 possible earthlike planets - a bonanza! There is, however, a catch: It is certain that the vast majority of these 87 are not real planets at all.
The algorithm is overly lenient in interpreting normal variations in starlight as being transits of small planets, and this was further complicated by abnormalities in the Kepler instrument itself. To explain this, it's necessary to say something of how Kepler operated.
During its operational lifetime, which ended earlier this year, Kepler would spend a quarter year (about three months) at a time staring very steadily at the same region of the sky. The light from Kepler's telescope falls onto an array of charge-coupled devices, or CCDs, like those in a digital camera.
At the end of each quarter, Kepler rotated its entire body 90 degrees, still staring at the same area of the sky, but with each star now observed by a different location on the surface of Kepler's grid of photosensitive CCDs for the new quarter. After that quarter, Kepler would rotate again, and so on for the four years, or 16 quarters, of Kepler's lifetime. This means that each particular star was observed by four different locations on Kepler's CCD grid, alternating in a cycle of four quarters, or one year.
The problem in this scheme is that certain locations of Kepler's CCD grid have proven to be slightly erratic. The data they collected are still useful, but with a little random noise, or jitter, added to the observations in ways that could make it seem like a small planet had transited in front of the star when in fact no transit at all had taken place. As a result, the same balky electronics could report bogus transits each quarter it observed the star, which therefore means at intervals of about a year.
If the bogus transits happened by chance to come at even intervals in time, this fooled the search algorithm into thinking it may have seen a small planet. What makes this particularly insidious is that these false reports of planets often seem to be earthlike in both size and temperature, including many of the 87 seemingly earthlike planets mentioned above.
That list of possible planets came from Kepler's first three years of observations, but fortunately, there is a saving grace - the fourth year. For three bogus transits to be spaced out evenly in time is not highly unlikely: The second bogus transit simply has to occur at the time halfway between the first and the third. But if the planet is real, then a fourth transit should occur about a year later, timed just as precisely as the first three. This is unlikely to occur if the first three were bogus. So by finding a fourth transit is a powerful reality check on that list of possible earthlike planets.
In my analysis of this data, I first studied the "noise" that Kepler recorded in observing each star to come up with an estimate of how likely it was that Kepler was registering some false transits associated with that star. Then, by focusing on a small list of stars which had a possible earthlike planet and seemed to escape the problem of Kepler's noisy electronics, I checked to see if their that planet showed an additional transit in the final year of observations. At least two, and perhaps five, showed signs of an additional and confirmatory transit.
Does this constitute the discovery of five truly earthlike planets around other stars? Not yet, because in any particular case there are other explanations for how real astrophysical events can mimic the transits of an earthlike planet. This figure (figure 2 lower image) shows three different cases that can look the same to Kepler. In case (a), we see an earthlike planet in the process of transiting its star, the case we are hoping to find.
In case (b), we see a large planet transiting a distant star, which just happens to be lined up right behind the closer (and therefore seemingly bigger) star that Kepler was looking at. In this figure, we can see what's happening, but Kepler doesn't obtain pictures with this level of detail; it only measures how the total amount of light varies over time.
When the large planet transits across the distant star, the loss of light is the same as if a small planet were transiting across the nearer star, so Kepler is effectively blind to the difference between these two cases. Case (c) shows another possibility, that a pair of binary stars perform a grazing eclipse of one another, hiding just enough of each other to block the same amount of light that an earth-sized planet would.
Very careful analysis of Kepler data may be able to tell when something like (b) or (c) is happening. In addition, even for those stars where (a) proves to be the case, the estimates of planet size and temperature are approximate, and some will prove to be larger, smaller, hotter, or cooler than the original estimate. It's left to future observations, using different telescopes and different methods, to examine these five or so cases to see which of these truly are what they seem they may be - planets with the same size and temperature as Earth.
Verification of these cases as earthlike planets, and further study of any of them that are, may prove to be a challenge as all of them are on the order of 2,000 light years away. But even without such verification, the number of possible earthlike planets resulting from this analysis provides an upper limit of how common such planets may be, and combined with estimates of how common cases (b) and (c) should be, can also provide a lower limit. This is not the final step in finding another place like home far off among the stars, but it's a critical step, and the one that Kepler was designed for.
Until now, Kepler's nearly continuous observations of over 150,000 stars have confirmed the existence of Earth-sized planets in the hot regions close to their star. Larger planets, some as small as one and a half times the Earth's diameter, have been found in the Habitable Zone, where the amount of heat they receive from their star may sustain earthlike temperatures.
In addition, earthlike planets have been found in the Habitable Zone of tiny, cool red dwarf stars, which may offer a more hazardous environment for life than with sunlike stars. But finding planets with the combination of earthlike size and earthlike temperature around sunlike stars, a major goal of the Kepler mission, has been elusive.
Part of the challenge is how difficult it is to spot a single transit by an Earth-sized planet, which blocks only 0.01% of its star's light, far less than larger planets block. This figure (figure 1 top image) shows the graph of three different stars' brightness as measured by Kepler during actual transits by extrasolar planets the size of Jupiter, Neptune, and Earth. Transits of larger planets stand out prominently, but transits by Earth-sized planets are so subtle, Kepler can only identify them when there is a series of many such dimming events occurring at the same regular interval.
Last December, the Kepler team released a list of 18,406 possible planets found by an algorithm that searched the first three years of Kepler data looking for series of transits.
If we define a size limit of no more than 1.25 times the Earth's diameter and a Habitable Zone where the amount of heating that a planet would receive is between that of Venus and Mars in our solar system, then this list contains 87 possible earthlike planets - a bonanza! There is, however, a catch: It is certain that the vast majority of these 87 are not real planets at all.
The algorithm is overly lenient in interpreting normal variations in starlight as being transits of small planets, and this was further complicated by abnormalities in the Kepler instrument itself. To explain this, it's necessary to say something of how Kepler operated.
During its operational lifetime, which ended earlier this year, Kepler would spend a quarter year (about three months) at a time staring very steadily at the same region of the sky. The light from Kepler's telescope falls onto an array of charge-coupled devices, or CCDs, like those in a digital camera.
At the end of each quarter, Kepler rotated its entire body 90 degrees, still staring at the same area of the sky, but with each star now observed by a different location on the surface of Kepler's grid of photosensitive CCDs for the new quarter. After that quarter, Kepler would rotate again, and so on for the four years, or 16 quarters, of Kepler's lifetime. This means that each particular star was observed by four different locations on Kepler's CCD grid, alternating in a cycle of four quarters, or one year.
The problem in this scheme is that certain locations of Kepler's CCD grid have proven to be slightly erratic. The data they collected are still useful, but with a little random noise, or jitter, added to the observations in ways that could make it seem like a small planet had transited in front of the star when in fact no transit at all had taken place. As a result, the same balky electronics could report bogus transits each quarter it observed the star, which therefore means at intervals of about a year.
If the bogus transits happened by chance to come at even intervals in time, this fooled the search algorithm into thinking it may have seen a small planet. What makes this particularly insidious is that these false reports of planets often seem to be earthlike in both size and temperature, including many of the 87 seemingly earthlike planets mentioned above.
That list of possible planets came from Kepler's first three years of observations, but fortunately, there is a saving grace - the fourth year. For three bogus transits to be spaced out evenly in time is not highly unlikely: The second bogus transit simply has to occur at the time halfway between the first and the third. But if the planet is real, then a fourth transit should occur about a year later, timed just as precisely as the first three. This is unlikely to occur if the first three were bogus. So by finding a fourth transit is a powerful reality check on that list of possible earthlike planets.
In my analysis of this data, I first studied the "noise" that Kepler recorded in observing each star to come up with an estimate of how likely it was that Kepler was registering some false transits associated with that star. Then, by focusing on a small list of stars which had a possible earthlike planet and seemed to escape the problem of Kepler's noisy electronics, I checked to see if their that planet showed an additional transit in the final year of observations. At least two, and perhaps five, showed signs of an additional and confirmatory transit.
Does this constitute the discovery of five truly earthlike planets around other stars? Not yet, because in any particular case there are other explanations for how real astrophysical events can mimic the transits of an earthlike planet. This figure (figure 2 lower image) shows three different cases that can look the same to Kepler. In case (a), we see an earthlike planet in the process of transiting its star, the case we are hoping to find.
In case (b), we see a large planet transiting a distant star, which just happens to be lined up right behind the closer (and therefore seemingly bigger) star that Kepler was looking at. In this figure, we can see what's happening, but Kepler doesn't obtain pictures with this level of detail; it only measures how the total amount of light varies over time.
When the large planet transits across the distant star, the loss of light is the same as if a small planet were transiting across the nearer star, so Kepler is effectively blind to the difference between these two cases. Case (c) shows another possibility, that a pair of binary stars perform a grazing eclipse of one another, hiding just enough of each other to block the same amount of light that an earth-sized planet would.
Very careful analysis of Kepler data may be able to tell when something like (b) or (c) is happening. In addition, even for those stars where (a) proves to be the case, the estimates of planet size and temperature are approximate, and some will prove to be larger, smaller, hotter, or cooler than the original estimate. It's left to future observations, using different telescopes and different methods, to examine these five or so cases to see which of these truly are what they seem they may be - planets with the same size and temperature as Earth.
Verification of these cases as earthlike planets, and further study of any of them that are, may prove to be a challenge as all of them are on the order of 2,000 light years away. But even without such verification, the number of possible earthlike planets resulting from this analysis provides an upper limit of how common such planets may be, and combined with estimates of how common cases (b) and (c) should be, can also provide a lower limit. This is not the final step in finding another place like home far off among the stars, but it's a critical step, and the one that Kepler was designed for.
No comments:
Post a Comment