Rife with hype, exoplanet study needs patience and refinement

Feb 18, 2014 by Morgan Kelly
Exoplanet transiting in front of its star. Princeton’s Adam Burrows argues against drawing too many conclusions about such distant objects with today’s technologies. Credit: ESA/C. Carreau

(Phys.org) —Imagine someone spent months researching new cities to call home using low-resolution images of unidentified skylines. The pictures were taken from several miles away with a camera intended for portraits, and at sunset. From these fuzzy snapshots, that person claims to know the city's air quality, the appearance of its buildings, and how often it rains.

This technique is similar to how scientists often characterize the atmosphere—including the presence of water and oxygen—of outside of Earth's , known as exoplanets, according to a review of exoplanet research published in the Proceedings of the National Academy of Sciences. A planet's atmosphere is the gateway to its identity, including how it was formed, how it developed and whether it can sustain life, stated Adam Burrows, author of the review and a Princeton University professor of astrophysical sciences. But the dominant methods for studying exoplanet atmospheres are not intended for objects as distant, dim and complex as planets trillions of miles from Earth, Burrows said. They were instead designed to study much closer or brighter objects, such as planets in Earth's solar system and stars.

Nonetheless, scientific reports and the popular media brim with excited depictions of Earth-like planets ripe for hosting life and other conclusions that are based on vague and incomplete data, Burrows wrote in the first in a planned series of essays that examine the current and future study of exoplanets. Despite many trumpeted results, few "hard facts" about exoplanet atmospheres have been collected since the first planet was detected in 1992, and most of these data are of "marginal utility."

The good news is that the past 20 years of study have brought a new generation of exoplanet researchers to the fore that is establishing new techniques, technologies and theories. As with any relatively new field of study, fully understanding exoplanets will require a lot of time, resources and patience, Burrows said.

"Exoplanet research is in a period of productive fermentation that implies we're doing something new that will indeed mature," Burrows said. "Our observations just aren't yet of a quality that is good enough to draw the conclusions we want to draw.

"There's a lot of hype in this subject, a lot of irrational exuberance. Popular media have characterized our understanding as better than it actually is," he said. "They've been able to generate excitement that creates a positive connection between the astrophysics community and the public at large, but it's important not to hype conclusions too much at this point."

The majority of data on exoplanet atmospheres come from low-resolution photometry, which captures the variation in light and radiation an object emits, Burrows reported. That information is used to determine a planet's orbit and radius, but its clouds, surface, and rotation, among other factors, can easily skew the results. Even newer techniques such as capturing planetary transits—which is when a planet passes in front of its star, and was lauded by Burrows as an unforeseen "game changer" when it comes to discovering new planets—can be thrown off by a thick atmosphere and rocky planet core.

All this means that reliable information about a planet can be scarce, so scientists attempt to wring ambitious details out of a few data points. "We have a few hard-won numbers and not the hundreds of numbers that we need," Burrows said. "We have in our minds that exoplanets are very complex because this is what we know about the planets in our solar system, but the data are not enough to constrain even a fraction of these conceptions."

Burrows emphasizes that astronomers need to acknowledge that they will never achieve a comprehensive understanding of exoplanets through the direct-observation, stationary methods inherited from the exploration of Earth's neighbors. He suggests that exoplanet researchers should acknowledge photometric interpretations as inherently flawed and ambiguous. Instead, the future of exoplanet study should focus on the more difficult but comprehensive method of spectrometry, wherein the physical properties of objects are gauged by the interaction of its surface and elemental features with light wavelengths, or spectra. Spectrometry has been used to determine the age and expansion of the universe.

Existing telescopes and satellites are likewise vestiges of pre-exoplanet observation. Burrows calls for a mix of small, medium and large initiatives that will allow the time and flexibility scientists need to develop tools to detect and analyze exoplanet spectra. He sees this as a challenge in a research environment that often puts quick-payback results over deliberate research and observation. Once scientists obtain high-quality spectral data, however, Burrows predicts, "many conclusions reached recently about exoplanet atmospheres will be overturned."

"The way we study planets out of the solar system has to be radically different because we can't 'go' to those planets with satellites or probes," Burrows said. "It's much more an observational science. We have to be detectives. We're trying to find clues and the best clues since the mid-19th century have been in spectra. It's the only means of understanding the atmosphere of these planets."

A longtime researcher, Burrows predicted the existence of "hot-Jupiter" planets—gas planets similar to Jupiter but orbiting very close to the parent star—in a paper in the journal Nature months before the first such planet, 51 Pegasi b, was discovered in 1995.

Explore further: Bright star reveals new exoplanet

More information: Burrows, Adam S. 2014. Spectra as windows into exoplanet atmospheres. Proceedings of the National Academy of Sciences. DOI: 10.1073/pnas.1304208111

add to favorites email to friend print save as pdf

Related Stories

Bright star reveals new exoplanet

Jan 22, 2014

An international team of astronomers at Stellar Astrophysics Centre in Aarhus, Denmark, have discovered a new exoplanet, christened "Kepler-410A b." The planet is about the size of Neptune and orbits the ...

Many small exoplanets found to be covered in gas

Jan 08, 2014

During its four-year mission, NASA's Kepler space telescope discovered thousands of "planetary candidates" in our Milky Way galaxy—the vast majority of which are almost certainly actual planets. But before scientists can ...

First planet found around solar twin in star cluster

Jan 15, 2014

Astronomers have used ESO's HARPS planet hunter in Chile, along with other telescopes around the world, to discover three planets orbiting stars in the cluster Messier 67. Although more than one thousand ...

What are hot Jupiters?

Feb 13, 2014

When astronomers first discovered other planets, they were completely unlike anything we've ever found in the Solar System. These first planets were known as "hot Jupiters", because they're giant planets ...

Possible water in the atmosphere of a super-Earth

Mar 05, 2012

A "super-Earth" is an exoplanet (a planet around another star) whose mass is between about two and ten Earth-masses. Planets larger than this are closer to Uranus and Neptune in size (and perhaps in other ...

Recommended for you

Quest for extraterrestrial life not over, experts say

19 hours ago

The discovery of an Earth-sized planet in the "habitable" zone of a distant star, though exciting, is still a long way from pointing to the existence of extraterrestrial life, experts said Friday. ...

Continents may be a key feature of Super-Earths

23 hours ago

Huge Earth-like planets that have both continents and oceans may be better at harboring extraterrestrial life than those that are water-only worlds. A new study gives hope for the possibility that many super-Earth ...

Exoplanets soon to gleam in the eye of NESSI

Apr 18, 2014

(Phys.org) —The New Mexico Exoplanet Spectroscopic Survey Instrument (NESSI) will soon get its first "taste" of exoplanets, helping astronomers decipher their chemical composition. Exoplanets are planets ...

User comments : 16

Adjust slider to filter visible comments by rank

Display comments: newest first

GSwift7
4 / 5 (4) Feb 19, 2014
Yay, finally a rational voice speaking up!

I've been saying this for a long time. You can't really just get one sample and interpret the results with much confidence, since there's no control to compare your results against. They can say just about anything they want, and there's no way to either verify or invalidate their interpretations. It's funny how much of this I've seen going on lately, and not just in the exoplanet field.

We really need to launch a follow-up mission to Kepler, and make sure it is able to stay active for 10 or 20 years, at least. If a planet has an orbit of 3 years of more, then Kepler will have only one data point for it, and even a dozen data points would be pushing it in terms of data integrity in most fields.

I think the story above makes a very good point regarding what our present knowledge is good for. While it is too early to attempt interpretation of data, that data tells us a lot about what kind of instruments we will need on next-gen missions
Torbjorn_Larsson_OM
5 / 5 (2) Feb 19, 2014
We can have legitimate reasons for skepticism. But the overview of Burrow's critique show him as tendentious.

- "From these fuzzy snapshots, that person claims to know the city's air quality, the appearance of its buildings, and how often it rains. This technique is similar to how scientists often characterize the atmosphere ...".

Yes, but with statistics. This is how people draw information out of images, from photos to radar, and with success - or the methods are abandoned.

- "But the dominant methods for studying exoplanet atmospheres are not intended for objects as distant, dim and complex".

Now they are, as is similar astronomical techniques that have been adopted from elsewhere. E.g. radar techniques showing us Venus's surface.

[tbctd]
Torbjorn_Larsson_OM
5 / 5 (1) Feb 19, 2014
[ctd]

- "That information is used to determine a planet's orbit and radius, but its clouds, surface, and rotation, among other factors, can easily skew the results."

Certainly. That is why replication is important.

Most intriguingly, it seems it works! A method precisely relying on atmosphere properties to derive planet radius and mass (so density can be derived), MassSpec, has recently shown the very same mass as RV. [ http://astrobites...ard-way/ ]

There is nothing like actual experiments to illuminate speculations such as this. I would say Burrow's larger concern is now known to be failed. His technical concern (photometry), well, he's the expert, point well taken. (MassSpec used transmission and emission spectroscopy both.)
Torbjorn_Larsson_OM
5 / 5 (1) Feb 19, 2014
@GS: Well, if you go by measurement theory you never "interpret", you test.

And sometimes you can use a single fact to reject hypotheses despite the presence of noise and systematic errors. For example, if you can observe an event from a fossil any number of times and with many details. The CMB fossil radiation or primordial nucleosynthesis in cosmology comes to mind, as does individual transitional fossils in biology or the presence of the impact-originated Moon in geology.

So I would be cautious about "interpreting" the idea that single samples can't be interpreted with much confidence, because we have found it is doable. The trick is to pinpoint when it doesn't work and/or when people makes erroneous claims from such messes.
Fleetfoot
5 / 5 (1) Feb 19, 2014
If a planet has an orbit of 3 years of more, then Kepler will have only one data point for it, and even a dozen data points would be pushing it in terms of data integrity in most fields.


Remember though that all the Kepler discoveries have only been "candidates", they all require confirmation by other telescopes. The problem is that only having one transit makes it harder to predict when to request telescope time for the follow up.
GSwift7
1 / 5 (1) Feb 20, 2014
Stupid 1000 character limt. My post came out sounding critical of the wrong things, I think, based on your responses.

Kepler results are fine for detecting candidates, and even confirming the ones with short orbital periods. The type of interpretations I'm talking about are more in regard to ground-based work that's been done by a few groups.

For example, there's been talk of 'diamond planets' based on the discovery of a rocky planet orbiting a carbon rich star. That's a fairly big leap of interpretation. There's another interpretation which sounds fishy where they're suggesting 'fluffy planets' with density that doesn't seem to make sense.

among other factors, can easily skew the results."

Certainly. That is why replication is important.


But without a control case, you can replicate a mistake in your assumptions. For example, we are using a factor to guess the size of planets based on the light blocked during transits. We don't have a baseline from which to verify that.
GSwift7
3.7 / 5 (3) Feb 20, 2014
continued:

Here are examples which might change that factor: How much light would the Earth have blocked following the collision that created the moon? Or, how about a planet with a dense ring system around it with enough obliquity so that we aren't seeing the ring edge-on? Here's another one: The assumption of size based on shadow depends on knowing how far the planet is from the star, and we guess that based on the orbital period. However, that assumes a circular orbit. We have no way of knowing if the orbit is eliptical and no way of knowing whether the transit happened at aphelion or perihelion, which would change the star-planet distance and thereby throw the size estimate off. A binary planet, or a planet with a moon of significant size would also throw off the size/density estimate. Would a water-rich terrestrial planet show hydrogen lines if a recent eruption has thrown dense volcanic ash into its upper atmosphere (think ancient olympus mons, for example)?
GSwift7
2 / 5 (2) Feb 20, 2014
I thought of one more big systematic error factor right after I hit the submit button:

The size of the parent star is only a rough estimate unless it is close enough to use parallax shift. Getting the size/distance of the parent star wrong by a few percent would throw off all your measurements of the system. For example, intervening gas or dust might make a star appear dimmer and therefore farther away than it really is.
antialias_physorg
5 / 5 (2) Feb 20, 2014
For example, we are using a factor to guess the size of planets based on the light blocked during transits. We don't have a baseline from which to verify that.

We do have Mercury and Venus (as well as, possibly, the observations from other probes we have sent further out into our solar system that have looked back. But I don't know if any of those have been purposely pointed at the sun to record transits of Earth, Mars, etc. )

Atmospheric conditions aren't likely to cause much of a difference in data as the atmosphere of planets is a rather thin film (for those with thick atmospheres it looks like all but the topmost parts block out the sun completely).
For example, intervening gas or dust might make a star appear dimmer and therefore farther away than it really is

Intevening gas would not cause redshift of the star's own absorption lines (redshift can be used for far away stars) nor its parallax (for close by stars).
GSwift7
3 / 5 (2) Feb 21, 2014
Intevening gas would not cause redshift of the star's own absorption lines


We can't do that for stars inside the milky way, which are the only ones we look at for exoplanets. We mainly use the aparent magnitude to judge distance to medium distance stars, afik, which are the bulk of exoplanet study candidates.

Yes, we have looked back at our own sun for transits of our inner planets, but that isn't a good analogy because the distance scale is much different and the instruments on board our deep space probes aren't the same as our exoplanet hunting instruments. You're still left with a relatively large margin of error (a couple percent can be a big deal) in several of the variables regarding characterization of exoplanets. The reason 1 or 2 percent can be so important is because the estimates feed off of each other. If the distance is off, then the size is off, which makes the density off, or the other way around.

We don't have a hot gas giant to sample here, so those are ??
GSwift7
3 / 5 (2) Feb 21, 2014
I agree about the atmosphere not being a factor in the shadow cast by a planet transit. Even a thick atmosphere like Venus' is nothing compared to the size of the whole planet. However, some other things we don't have examples of in our solar system might change the amount of light blocked. For example, the relative size of the star versus the planet combined with the distance between them and the planet's albedo at different wavelengths will change the signature of transits. A very reflective planet will shine around its edges and thereby appear smaller than it really is. If the planet is much smaller than its star, or closer to it, that effect will increase due to simple geometry. So, if you're off on the distance from Earth to the star, then you start throwing all those other factors off by a bit too (in the same direction) and they add up to a bigger and bigger error in your final estimates. I actually think something like that is highly likely in the case of their "fluffy planets"
TheGhostofOtto1923
1 / 5 (1) Feb 21, 2014
However, that assumes a circular orbit. We have no way of knowing if the orbit is eliptical and no way of knowing whether the transit happened at aphelion or perihelion
Uh no it doesn't. Scientists are fully aware of the prevalence of elliptical orbits and include this in their analysis.
For example, intervening gas or dust might make a star appear dimmer and therefore farther away than it really is
Again, the experts who are doing this work are well aware of your ad hoc caveats and include all of them and much much more in their analyses.
The reason 1 or 2 percent can be so important is because the estimates feed off of each other
These are SCIENTISTS. They know these things.
TheGhostofOtto1923
1 / 5 (1) Feb 21, 2014
there's been talk of 'diamond planets'
The idea was that if such a planet was composed of carbon, conditions would cause that carbon to be in crystalline form. This extrapolation was not the result of direct observation.

But scientists have since revised their analysis.

"carbon – the chemical element diamonds are made of – appears to be less abundant in relation to oxygen in the planet's host star..."

-This does not exclude the possibility of crystalline carbon forming under similar conditions.

"a layer of carbon in the form of graphite, possibly with a kilometers-thick substratum of diamond if there is sufficient pressure. During volcanic eruptions, it is possible that diamonds from the interior could come up to the surface, resulting in mountains of diamonds and silicon carbides."
GSwift7
5 / 5 (1) Feb 21, 2014
Otto:

Oh yeah? The guy in the article here disagrees with you, and he's one of the leading experts in the field. I'm merely pointing out specific holes in our emperical knowledge. This isn't some kind of State secret, and it's not really debated by anyone, which makes it puzzling when you see sensationalistic media releases like the ones this guy is complaining about.
TheGhostofOtto1923
1 / 5 (2) Feb 21, 2014
The guy in the article here disagrees with you, and he's one of the leading experts in the field. I'm merely pointing out specific holes in our emperical knowledge
No, he's saying that despite being aware of ALL the things you mentioned as well as everything else which goes into their analysis, these scientists are still pushing the envelope.

YOU dont have the education or experience to be pointing out holes. Sorry.

Perhaps hes right. Or perhaps he has a little too much time on his hands. Perhaps you have too much time on your hands.

For instance what could possibly make you think that astrophysicists wouldnt know about elliptical orbits??

"Celestial mechanics is the branch of astronomy that deals with the motions of celestial objects. The field applies principles of physics, historically classical mechanics, to astronomical objects such as stars and planets to produce ephemeris data"

-How many courses do you think they have to ace on this subject? How many have you aced?
GSwift7
not rated yet Feb 24, 2014
YOU dont have the education or experience to be pointing out holes. Sorry


I don't need a specialized degree in planetology to understand basic science techniques. The problems I mentioned are fairly easy to understand, and I could point out analogous examples of the same types of problems in any other field of observational science. Anyone in commercial manufacturing must deal with statistical analysis and sampling errors, as well as unknown cause and effect relationships, for example. That part of it isn't rocket science, and applying those principles to planetology isn't a big leap.

what could possibly make you think that astrophysicists wouldnt know about elliptical orbits??


That's not what I suggested. The truth is that the transit method is extremely poor at telling you whether the orbit is eliptical or not. All it tells you is the period of the orbit, and maybe the transit time, if your data is good enough. You can't get eccentricity with only that data.

More news stories

Impact glass stores biodata for millions of years

(Phys.org) —Bits of plant life encapsulated in molten glass by asteroid and comet impacts millions of years ago give geologists information about climate and life forms on the ancient Earth. Scientists ...

The importance of plumes

The Hubble Space Telescope is famous for finding black holes. It can pick out thousands of galaxies in a patch of sky the size of a thumbprint. The most powerful space telescope ever built, the Hubble provided ...

Continents may be a key feature of Super-Earths

Huge Earth-like planets that have both continents and oceans may be better at harboring extraterrestrial life than those that are water-only worlds. A new study gives hope for the possibility that many super-Earth ...

Airbnb rental site raises $450 mn

Online lodging listings website Airbnb inked a $450 million funding deal with investors led by TPG, a source close to the matter said Friday.

Health care site flagged in Heartbleed review

People with accounts on the enrollment website for President Barack Obama's signature health care law are being told to change their passwords following an administration-wide review of the government's vulnerability to the ...