Archive

Archive for June, 2006

flowchart

June 29th, 2006 2 comments

cat's eye

One of the systemic project’s most important goals is to reach the point where we can have genuinely realistic and aesthetically satisfying simulated images and animations of extrasolar planets. This will serve the scientific purpose of allowing us to get better comparisons with infrared observational data, and will ultimately allow us to embark on vicarious missions to the new-found worlds of our Galaxy.

In the interim, there’s a lot of coding and computing to do.

As described in this post from earlier this month, I’m advising UCSC Physics graduate student Jonathan Langton on a Ph.D. thesis geared to simulating the atmospheres of irradiated extrasolar planets. Jonathan finally graded his way through a horrific stack of lab reports and final exams, and has now been able to put full focus on the research. Progress is evident in a heavy stream of e-mails containing increasingly detailed animations.

Jonathan is using a numerical technique known as the pseudo-spectral method to do his simulations. The key idea is that the flow pattern on the surface of the simulated planetary sphere can be broken down into a superposition of Fourier modes. For example, as one moves around the planet, the longitudinal variations in the flow can be described in terms of a superposition of sinusoidal patterns. Sinusoids have analytically computable derivatives, which allow one to make a highly accurate representation of changes in the flow without resorting to a cripplingly large amount of computation.

Spectral methods have their drawbacks, however, in the form of high-frequency numerical noise. This noise was evident in the earlier simulations in the form of transient ribbed patterns within the flow. Over the past few days, Jonathan has designed an elegant filtering scheme which seems to be working very well in supressing these spurious features without killing the actual structures in the flow.

planet after 2 rotation periods

The snapshot above is from a test-calculation that implements the new filtering scheme. It’s part of an animation that simulates the development of an initially random vortical flow on the surface of a planet with the radius, mass, and rotation period of HD 209458 b. (Potential vorticity is the quantity plotted, resolution is 256×128, and the simulation runs for 5 rotational periods).

frames from the animation

Here is a link to Jonathan’s latest (7MB) animation. It’s hot off the computer.

Categories: worlds Tags:

Repo Man

June 27th, 2006 3 comments

Everything takes longer than you think it’s going to take.

I thought it would be a relatively straightforward task to collect and assemble all of the published radial velocity data sets together in a uniform format. Turns out (as is often the case) that I was overly optimistic. It’s been a major effort to get an authoritative radial velocity catalog into shape. Eugenio, however, has been extremely persistent and methodical, and the job is now essentially done. Datasets for 155 stars accompanied by published planets are now available on (1) the downloadable console, (2) the web-based console, and (3) on the systemic back-end. Many of these data sets are now available in ASCII format for the first time; Eugenio made extensive use of the Dexter applet to extract data from papers in which the radial velocities have been hitherto published only as plotted points.

As far as systems with published planets that are not on the console go, we’re definitely scraping the bottom of the barrel. This morning, Eugenio sent me an update on where he’s at with the last dregs. Basically, there are a dozen planet-bearing stars that still need to be added to the console. In most of these cases, we either can’t find any listing of data, or the data is available only in the form of a phased plot that can’t be disentangled:

0. HD114762: the two references I was able to find have (or appear to have) no uncertainties since the “planet” is likely a brown dwarf, I’ll still skip this one for now.

1. HD41004: hierarchical quad system: A(K star)-2.5 M_J/B(M star)-BD. Swiss give table of velocities for both A and B but no uncertainties.

2. Tau Bootes: This still looks like a hopeless cause.

3. GL86: phased velocities only.

4. HD11964: missing data?

5. HD122430: missing data?

6. HD196885: missing data?

7. HD34445: missing data?

8. HD59686: Only announced at American Astronomical Society Meeting; Still listed as Mitchell et al. 2004, ApJ, submitted.

9. HD73256: phased velocities only (do not confuse with HD 73526).

10. HD89307: missing data?

11. Tres-1: Phased velocities only.

Without a doubt, there are some easy as-yet unannounced and as-yet unpublished planets ripe for the picking off of the console menu. Two months ago, I wrote a series of posts showing that 51 Peg almost certainly has a second Saturn-mass planet in a habitable orbit. This planet was uncovered after only a few minutes of work on the console. This afternoon, Eugenio and I looked at four or five data-sets (basically at random) and found a nice planet candidate that we’re planning to write up in one of this week’s posts.

There are more than 6 billion people on Earth, and only a handful of them have discovered a planet. Here’s your chance.

Categories: detection Tags:

Alpha and Proxima

June 24th, 2006 1 comment

Proxima and Alpha Centauri are the Sun’s closest stellar neighbors. As they drift through the void, a mere twenty four trillion miles beneath the Earth, they exert a special fascination. Do worlds orbit these stars? Will we ever reach them? Will anyone ever stand on the surface of Alpha Centauri A “b” to witness the double sunrises that occur every 39.5-odd years?

starfield showing proxima centauri

Tiny Proxima lies a mere 15,000 AU from the Alpha Centauri AB binary pair, and moves with them through the Galaxy in a very similar direction and with a nearly identical speed. The likelihood of such a stellar configuration occurring purely by chance is less than one in a million, and based on this incredibly improbable arrangement, it has been suspected (since 1917) that the three stars constitute a bound triple system.

It’s too bad our solar system doesn’t have a companion like Proxima. If the nearest red dwarf lay 15,000 AU instead of 260,000 AU away, it would shine with the 3rd magnitude. It would be easily visible to the naked eye, and its parallax would amount to 1/120th the diameter of the full Moon. If Proxima belonged to us, rather than to Alpha Centauri, then the distances to the external stars would have likely been first measured directly by someone like Robert Hooke in the 1600s rather than Friedrich Bessel in the 1800s. We would now be avidly searching Proxima for possible terrestrial planets, and the prospects for interstellar travel would not seem quite so daunting.

Every other academic year, I teach a graduate course on astrophysical dynamics at UCSC, and one of the requirements for completion of the class is a piece of original research. During the Summer prior to the start of the class, I design a set of projects, and then we collaborate to see them through to completion.

The Alpha-Proxima Centauri system is an excellent source of projects. All three stars are extremely well characterized, and it’s interesting to look at the sorts of orbits that the configuration can support. In the Fall 2004 course, first-year UCSC graduate student Jeremy Wertheimer started to work on the following problem:

Let’s imagine that we want to send a probe to the Alpha-Proxima Centauri system, and for the sake of concreteness, let’s assume that the probe can be accelerated to a large speed by a multi-stage rocket, but that it carries no fuel of its own for the purposes of orbital insertion. Using only the principles of gravity assist and de-assist [see diagram below], and employing the gravitational fields provided by the three stars, what is the largest speed with which the probe can approach the system, and be brought into a bound orbit about any of the three stars? This maximum speed of approach serves to define a characteristic travel time to the system. How long is this time?

[Note that this is a dynamics problem falling under the general topic of multi-parameter minimization, and is not a mission proposal! There are certainly better, more effective ways to visit Alpha Centauri.]

diagram illustrating gravitational assist

The strategy for solving the problem is as follows: First, set up a model of the orbits of the three stars about their common center of mass. Then, define a “population” of trajectories involving possible approaches (parameterized by velocity, impact parameter, and angle of attack). Use a genetic algorithm to breed promising trajectories, and after
many generations, arrive at one that is (hopefully) near-optimal.

As soon as we set to work on the problem, we found a remarkable result in the literature. In order to optimize the trajectory, we needed to know the most accurate available orbital parameters for Alpha Centauri A, B, and Proxima. To our surprise, we discovered that the most recent papers which study the system dynamics Anosova et al. 1994, and Matthews & Gilmore 1993, both suggest that Proxima is not gravitationally bound to Alpha Centaui AB. The results in the literature imply that the three stars are independent of one another, and just happen to be experiencing a close encounter while moving in the same general direction, despite the approximately million-to-one odds.

We were astonished by this result. It just didn’t seem to make sense. In addition to having the same kinematics, Proxima also seems to have an age and metallicity consistent with those of Alpha Centauri AB. Furthermore, in order to solve our minimum travel time problem, we needed to know whether the literature result is correct. The two extant papers, written in 1993 and 1994, were published prior to the release of the highly accurate Hipparcos data. Surprisingly, as far as we can tell, nobody has attempted to use the modern measurements to see whether Proxima really is unbound from the AB pair.

We therefore realigned Jeremy’s research project to provide an updated anlaysis of Proxima’s dynamical situation. Is it bound to Alpha or not? In an upcoming post, I’ll talk about what we’ve found.

Categories: worlds Tags:

Zoom

June 23rd, 2006 4 comments

image formed by a converging lens

Remember the scene in Blade Runner in which Deckard successively zooms and enhances a digitized photograph found at a crime scene? In 1982, it seemed to epitomize the fashionably sleek high tech, and it left a strong impression on me.

In this post from last February, I wrote about the protostellar disks in Orion that were imaged by the Hubble Space Telescope in the mid-1990s. One of these disks has achieved nearly iconic status (at least among those of us who give talks on planet formation). The following image shows it viewed edge-on and in silhouette against a background of glowing nebular gas. Only a faint smudge of red hints at the central star embedded within the disk.

A protostellar disk in the Orion Star-Forming Region

This disk is roughly 17 times larger than the orbit of Neptune. It’s also considerably larger than the orbits of 2003 UB-313 and Sedna, which I’ve integrated and placed on top of the image for comparison.

sedna's orbit

The Hubble press release that accompanies the disk image highlights a number of different protostellar disks, or “proplyds”, many of which are being strongly photoevaporated by radiation from the nearby high-mass stars of the Trapezium cluster. This region is faintly visible to the naked eye as the unresolved middle “star” of Orion’s sword:

orion showing trapezium

A growing body of evidence suggests that our own solar system may well have formed in a similarly disruptive environment. Analysis of meteorites shows evidence that radioactive atoms with short half-lives (in particular, Aluminum 26) freshly ejected from a nearby supernova in the birth cluster may have been incorporated into our solar system’s protostellar disk. The scattered orbits of bodies such as Sedna also hint that the solar system may have had a close encounter with another star at an early time in its history. Close encounters only occur in a dense stellar environment.

On the Hubble press release page, there is a 22.97 MB Tif image of the entire Trapezium region. When the full image is displayed at laptop-screen resolution, it isn’t clear where the protostellar disks actually are. If, however, you slog through the full download and open the image in a program like Photoshop, then, like Deckard, you can zoom in with successively higher resolution to find the disks shown in the press release. The resolution in the 23 MB Tif image is not the full resolution provided by the actual mosaic of images, but it’s high enough to enable discovery of a lot of detail. A leisurely exploration of the image with pan and the zoom controls gives an amazing sense of the overall structure of the stellar nursery. The three-dimensionality of the cluster is easier to visualize. You can sense that the disk is suspended in empty space, in a slow, arcing free-fall through the cluster, making it somehow easier to grasp that this is an image of new worlds in the process of creation.

full view

zoom 1 view

zoom 2 view

Categories: detection Tags:

A Hot Jupiter Simulation

June 18th, 2006 Comments off

eggs in an egg carton

Last week, we had a one-day seminar on planets and planet formation at UC Santa Cruz that brought together researchers from both UCSC and NASA Ames. One of the talks was by Jonathan Fortney, who is currently a post-doctoral researcher in the Planetary Systems Branch of the Space Science Division at NASA Ames.

Fortney and his NASA Ames collaborator Mark Marley have a state-of-the-art radiative transfer code which can compute the emergent (and reflected) spectrum from a hot Jupiter. (See this recent post.) They’ve recently applied their code to compute how the intrinsic radiation from the flow pattern on the surface of the planet would look if you could resolve it with a pair of night-vision goggles. Jonathan writes:

Here’s an MPEG of the full 360 orbit of HD 209458b, in 36 10-degree increments. This is as seen from Earth. It’s the Cooper & Showman (2006) dynamical simulation, run through our radiative transfer solver.

Red is 5 microns, green is 3.3 microns, and blue is 2.2 microns. The Cooper and Showman model predicts a day side that is very similar to a blackbody, leading to a whitish appearance. On the night side, which is fairly cool, strong methane absorption knocks out the blue and green, leaving only red. I have artificially pumped up the red on the night side so that you can actually see it on the monitor. If you don’t, it’s a dark red which is hard to see compared to black–the night side has little flux compared to bright (hot) day.

Here’s a sequence of frames from the movie:

frames from Fortney's HD209458b animation

The animation draws on calculations described by Fortney, J. J., et al., 2006, “The Influence of Atmospheric Dynamics on the Infrared Spectra and Light Curves of Hot Jupiters”, which has been submitted to the Astrophysical Journal.

This is a big step forward for the “computational imaging” of extrasolar planets, and I’m really excited about the future directions that Fortney is planning to take these calculations. For starters, it will be very interesting to see the movie with the reflected light component added in. It will also be cool to place the point of view above a particular spot on the planet and animate the time-dependant flow pattern (the above movie rotates a single snapshot model of the planet, but it does not show the actual time evolution that is computed in Cooper and Showman’s hydrodynamical simulations). Animations of the time-dependant flow will start to bring exoplanets into the territory covered by the cloud-pattern movies that the Voyager and Cassini probes radioed to Earth as they flew past Jupiter. Finally, by using John Moore’s integrating sphere to produce the actual visible colors corresponding to individual computed spectra, it will also be possible to produce true visible light (rather than night-vision-goggle infrared) animations of the simulated surface of HD 209458 b and other hot Jupiters. (In particular, HD 80606!)

Categories: worlds Tags:

Shallow Water

June 16th, 2006 3 comments

shallow water

Regular oklo readers all know that HD 209458 b is a lot bigger than it’s supposed to be.

In a previous post, we saw that the theoretical models that provide reasonable matches for the other 9 transiting planets predict that HD 209458 b’s radius should be slightly larger than Jupiter’s radius. The observations, on the other hand, make it clear that the planet is actually has a diameter about 1.35 times larger than Jupiter. HD 209458 b is by far the best-studied exoplanet, so it’s of more than passing interest to understand why it’s so large.

There’s general agreement that HD 209458 b must be privately tapping an unusual source of internal heat. Somehow, a lot of extra energy is being generated in the planetary interior. The surplus heat allows the planet to maintain an expanded outer envelope, and hence endows the planet with a larger overall size. The big question is: what is the anomalous extra heat source? A few years ago, I was enthusiastic about the idea that there might be a second companion planet that is gravitationally perturbing HD 209458 b, forcing it to maintain a slightly eccentric orbit. This eccentricity would be continually damped as a result of tidal interactions with the parent star, which would generate a sufficient amount of interior heating. Such a state of affairs is analogous to the heating of the inner Jovian satellites. The heat generated by tidal friction lends Io its off-the-hook volcanism, and maintains Europa as the Astrobiology poster world.

Unfortunately, however, the perturbing companion model no longer seems to be a viable explanation of HD 209458 b’s large size. That is, if you use the systemic console to fit to the HD 209458 b radial velocities, you’ll find that there is very little latitude for inserting significant extra planets. Try it and see, and upload your fits to the systemic back-end.

Josh Winn (MIT) and Matthew Holman (Harvard-Smithsonian CfA) have written a paper that presents an interesting hypothesis for resolving the HD 209458 b radius dilemma. Winn and Holman propose that the planet is caught in a so-called Cassini state, which is a resonance between spin precession and orbital precession. In a future post, I’ll give a heuristic discussion of the dynamics of how this situation can arise, and how the Cassini states work, but in short, if HD 209458b is trapped in the “Cassini state 2″, then its spin axis will lie almost in the orbital plane. Like all hot Jupiters, the planet will spin once per orbit, but it will literally be lying on its side as it orbits the parent star. A synchronous planet in Cassini state 2 will experience a large amount of tidal heating, even in the complete absence of any other planets in the system.

I like the Winn-Holman hypothesis because it’s potentially testable. If the planet is in Cassini state 2, then the pattern of illumination on the surface, and hence the time-dependant global infrared signature, will be very different than if it is locked into the standard upright configuration. In the standard scenario, a hot Jupiter has a fixed substellar point on its equator that does not wander significantly as the planet executes its orbit. One hemisphere of the planet is in perpetual day, while the other hemisphere experiences an endless night. Hydrodynamic calculations by James Cho and his collaborators (link), and by Adam Showman and his students (link), suggest that hot Jupiters should have a single strong equatorial jet that advects heat from the hot dayside to the cool night side. The oklo splash image has been adapted from Cho’s calculations, and shows this jet in action (see this post for more discussion).

I’ve been advising UCSC Physics graduate student Jonathan Langton, who has recently begun a study of what the flow pattern on a hot Jupiter should look like if the planet is caught in Cassini state 2. If the planet’s rotation axis lies in the orbital plane, and if the planet spins on its axis once per orbit, then the play of light and shadow across the planetary orb has a pattern that is totally unlike our seasons here on Earth. At the north and south poles, of a spin-synchronous Cassini-state-2 planet, the parent star rises, passes directly overhead, and then sets once per orbit. At one special spot on the equator, on the other hand, the star is always visible, and additionally passes directly overhead once per orbit. At the opposite spot on the equator (which we’ll call the anti-stellar point), the star never fully rises, but rather peeks half of its diameter above opposite horizons once per orbit.

antistellar point

Jonathan has made two short .avi format animations that help to illustrate the situation. In the first animation, we hover above the point on the equator that receives maximum illumination. In the second animation, we hover above the point on the equator that receives the least illumination. The mythology on such a world would likely be pretty interesting.

When a spin-synchronous planet is illuminated in this bizarre manner, the flow pattern on its surface should be very different than the flow pattern that would occur if the planet is in the standard upright configuration. Jonathan has finished a preliminary set of simulations using the so-called shallow water approximation which indicate that this is indeed the case. (The shallow water approximation is a 2-dimensional method for simulating atmospheric dynamics on the surface of the planet under the assumption that the depth of the fluid is much smaller than the horizontal scales of interest. Use of this approximation doesn’t require us to assume that the red-hot Jupiter is actually covered with water!)

Here are two of Jonathan’s .avi format animations that show the (still very) preliminary results. The first animation [11 MB, modem users watch out!] shows the evolution of the temperature distribution (on the anti-stellar hemisphere) for a planet in Cassini state 2. Here’s a snapshot at a particular moment in time:

temperature snapshot

The second animation [12 MB] shows the distribution of vorticity across the planet surface. The vorticity at a particular spot in a fluid flow can be thought of as the ability of the flow to cause a tiny imaginary paddle-wheel to spin. Here’s a snapshot from the animation. The high-vorticity orange structure is a giant fiery hurricane-like storm on the surface of the planet:

vorticity snapshot

Categories: worlds Tags:

Illustrations

June 11th, 2006 3 comments

fan palm

I think readers sometimes wonder why oklo.org posts about extrasolar planets tend to be illustrated with seemingly random photographs from my house, my yard, and my neighborhood that bear (at best) a distant relation to the topic at hand. I’m liable to get written up on charges of pseudo-artistic pretentiousness for attempting to run the Jones Soda of Astronomy.

We’re taking this approach for several reasons. The oklo blog is designed to recruit users for the systemic collaboration, and and I’m paying the ISP out-of-pocket. An idiosyncratic format makes it easier to keep the posts flowing in the midst of the chronically never caught up academic routine of teaching, research, qualifying exams, topic defenses, homework grading, proposal writing, committee meetings, undergraduate theses, graduate advising, editing, visiting speaker hosting, etc.

Another reason is to drive home my belief that the real interest, the real fascination with extrasolar planets will ultimately lie in their tiny ephemeral details. It’s one thing to gain an abstract, theoretical understanding of the growth of planetary cores through the accretion of small bodies — it’s quite another to see a one-kilometer bolide succumb to the fiery tendrils of fluid instability as it slams into the toxic atmospheric murk of an unsettled four Earth-mass world. It’s one thing to know Neptune’s orbital eccentricity to five significant figures, it’s quite another to swoop in close to see the flow of feathered cirrus outline the turbulent core of the great dark spot.

I can’t get over the fact that I can photograph the subtly intricate details of a habitable planet available in my own backyard, and later that afternoon have them transmitted digitally across the globe. Just think, if we had an autonomous lander with a 5-megapixel camera engaged in a small-scale survey of a hectare-sized region of an Earth-mass terrestrial planet in the habitable zone of any G2V star other than the Sun, then this site would be getting a lot more than 400 visitors per day.

Finally, in resorting to the use of photographs of familiar objects to illustrate unfamiliar things, I want to underscore the urgent need for scientifically correct visualizations of extrasolar planets.

New planet-related discoveries are the subject of numerous NASA and NSF press releases and press conferences, and because these dicoveries generally report information obtained by indirect observational techniques, there’s a need for illustrations to accompany the releases “to capture the public interest”.

During the last ten years, these images conveying scientific results have generally been supplied by artist’s impressions. On occasion, some of these have veered toward the bizarre, the lurid, and the just plain wrong. Consider, for example, the above two illustrations of HD 209458 b. Both appeared in fairly recent HST press releases, and both are riddled with profound misconceptions. First look at that painting on the left. At the time when the image was most recently released, it was well-known that satellite orbits around the planet are highly dynamically unstable (e.g. Barnes and O’Brien 2002). The shadowing of the planet only makes sense if the star is much smaller than the planet and is somehow orbiting just above the planetary surface. The number of cloud bands and zones indicate that the planet is rotating with a ~12 hour period (like Jupiter) whereas in reality, the planet must be spin-synchronous with its P=3.5257 day orbit, and should probably have a single prominent equatorial jet. The panel on the right, which attempts to show that the planet is surrounded by an optically thin, yet Lyman-alpha-absorbing hydrogen cloud, is dramatically inconsistent with the laws of perspective, illumination, and radiative transfer. Literally billions of dollars are being spent on extrasolar planets. Surely, at the pinnacle of our dispatches, we can do better than this.

Genuinely realistic visualizations of extrasolar planets that incorporate all known information in a self-consistent way are going to become an oklo.org rallying cry over the coming year. We won’t be flying to Upsilon Andromedae any time soon. Radial velocities, photometric light curves, stellar and planetary models, dynamical integrators, radiative transfer routines, hydrodynamic codes, and Maya are what we have to work with. If it is the destiny of the extrasolar planets to truly inspire, then we must strive to see them as they really are.

Categories: worlds Tags:

Transit Fever

June 8th, 2006 1 comment

peppercorn on a blood orange

Literally every astronomical worker, amateur and professional alike, who has carried out a serious photometric search for planetary transits is familiar with the symptoms of at least one of the two common strains of transit fever.

In the egressia strain, one observes a photometric time-series in which a transit seems to be ending just as the observations are starting:

egressia

The ingressia strain has different symptoms, but is equally infectious. Near the end (or sometimes near the middle of an observing session, it appears from the time-series photometry that a planet is entering transit. In the most common form of the syndrome, the star generally either descends into the murk of high air-mass or is overtaken by dawn before the transit has finished:

ingressia

A third, somewhat rarer variety of the fever, known as rossiteria, has also been described. This strain is most commonly contracted by theorists; one finds radial velocities in the literature taken during a transit window which seem to show clear evidence of the Rossiter-McLaughlin effect:

rossiteria

I came down with my first serious case of transit fever (later diagnosed as egressia with complications due to rossiteria) nearly four years ago. The symptoms were brought on by HD 217107, the first candidate planet-bearing star observed by the transitsearch.org network. On the night of August 6th, 2002, photometric data was sent by both an observer in Pleasanton California, and from the KAIT automatic telescope at Lick Observatory. The fits to the radial velocity data indicated that the HD 217107 b transit window was scheduled to begin at 2:40 am PDT, and amazingly, both data sets showed a photometric dip right at the predicted time. To seemingly clinch the case, Debra Fisher also used the Lick 3-meter telescope to obtain five radial velocity measurements of HD217107 during and before the predicted transit ingress. When the spectra were analyzed, the velocities came back with a pattern consistent with the expected Rossiter-McLaughlin effect! [The data is available at this webpage]

To say that I was excited was an understatement. It was my first time out. I had no natural resistance. Tim Castellano, Debra, and I all came down with a full-blown case of transit fever. The period of HD 217107b is 7.127 days, which means that successive transits are spaced one week and 3 hours apart. Tim and I were on the verge of flying with a Meade LX-200 to Hawaii to observe the Aug. 13th transit from the parking lot of the Keck Observatory (That plan, fortunately, was canceled by a combination of high ticket prices and Joe Miller, the cooler-headed then-director of UCO/Lick).

By mid-September, observers in the Canary islands obtained a data set which clearly shows that HD 217107b does not transit. Huge disappointment. At that time, HD 209458 b was still the only known transiting extrasolar planet, and so the second detection would have been a very big deal.

With hindsight, having been innoculated against the transit fever, it’s clear that the transit interpretation was ambiguous in all three of our data sets. For example, in the case of the radial velocities from the Lick 3-meter, a new CCD detector for the spectrograph had just been installed, and so the overall zero point of the velocities relative to the predicted radial velocity curve was a free parameter. If the points are all moved down by 10-15 m/s, then the transit feature turns into ordinary scatter in the data. Likewise, with the photometric data, it was clear in retrospect that systematic effects were at work.

The transit detection problem is tough in part because it’s extraordinarily easy for systematic effects to seemingly conspire to produce an apparent signal. I would not feel confident in announcing a transit until I’ve seen multiple full-transit light curves. On the other hand, though, the false alarms play an important role. They get observers out on the sky, and spur the collection of enough data to truly rule out an event. This certainly wound up being the case for GJ 876, HD 168746, and a number of other candidates.

An early case of transit fever was contracted by U. J. J. LeVerrier, the Nineteenth-century French mathematician famous for the dynamical calculations that led to the prediction and subsequent discovery of Neptune. After the Neptune discovery, LeVerrier turned his attention to explaining the precession of Mercury’s perihelion, and found that the effect could be explained by the presence of a small intra-Mercurial planet that he named Vulcan. After a thorough literature search, LeVerrier unearthed five separate observations of the solar-disk transits by this planet, all at times consistent with predictions. In the following 1877 communication to the Monthly Notices of the Royal Astronomical Society, he’s basically saying, how could all those observations (which agree with theory, no less) possibly be wrong?

As everyone now knows, LeVerrier’s Vulcan doesn’t exist. The 43 extra seconds of Mercurian perihelion precession that had bothered LeVerrier so severely are explained by modifications to classical Newtonian gravity by Einstein’s general theory of relativity.

Categories: detection Tags:

Radius anomalies?

June 3rd, 2006 Comments off

Tristan Guillot and his colleagues have just published a paper, “A correlation between the heavy element content of transiting extrasolar planets and the metallicity of their parent stars” which explores an interesting new hypothesis for resolving the size problem for transiting hot Jupiters.

Readers of oklo.org are well aware that our theoretical understanding of the radii of hot Jupiters isn’t all it could be. For example, the transiting planets TrES-1 and HD 209458 b have very similar masses and surface temperatures, and yet HD 209458 b has a radius that is roughly 25% larger than TrES-1′s. In a previous post (see also this post), we outlined some of the hypotheses that might explain this discrepancy in radii.

Guillot et al’s idea is that the mass of a planetary core is a very steep function of stellar metallicity. That is, doubling the metallicity of the parent star leads to a 5-10 fold increase in the amount of mass contained in the cores of any short-period planets in orbit around the star. Larger core masses lead to smaller overall planetary radii at given mass, and so, in the Guillot et al. picture, planets orbiting metal-rich stars will, in general, be considerably smaller than planets of equal mass orbiting metal poor stars.

They present the following graph to support their hypothesis. It shows the difference in observed planetary size from the baseline theoretical expectation on the y-axis, and the parent star metallicity on the x-axis. They note (and the eye notes) that there is a trend in the diagram; planets orbiting metal-rich stars (as exemplified dramatically by HD 149026b) tend to be smaller than predicted, and planets orbiting stars of near-solar metallicity (e.g. HD 209458b tend to be larger than predicted.

Figure adapted from Guillot et al. 2006.

But is this correlation really present? To get a qualitative sense of whether it is or not, I took the planetary radius and parent star metallicity values for the 9 transiting planets in the Guillot et al plot and redrew them from their implied uncertainty distributions to make alternate, statistically equivalent versions of the plot in their paper. I also made control plots in which I assumed that there is no underlying radius-metallicity effect, only noise from measurement uncertainty. I made four plots of the first variety, and four plots of the second control variety. They are shown below. Can you identify which plots are the control plots?

Categories: detection Tags:

Pantone 272

June 1st, 2006 Comments off

I’ve often wondered what the dayside of a Hot Jupiter would look like (with the dark, wraparound shades on, of course).

Jonathan Fortney and his collaborators at NASA Ames Research Center have been making sophisticated calculations to determine the atmospheric structures of giant planets. As a product of this research, they can compute a prediction of what the spectrum of the light coming from a short-period giant planet should be. For example, for HD 209458 b, they get the following distribution of light coming off of the planet:

If we look at the dayside (substellar point) of the planet, the total distribution of light that we would see in this model is given by the solid black curve. This total light is a sum of the light that the planet reflects (the blue line) and the light that is actually generated by the hot planet itself (the red line). As one would expect, the visible-light dayside appearance of the planet is dominated by the reflected light. The nightside image is considerably les bright, and its somewhat battered thermal spectrum is a bit more magenta than the Oklo splash image, (which was generated with a black-body color map).

Molecular absorption due largely to water and methane selectively removes yellow and green light from the reflected optical spectrum. This suggests that the hot Jupiter daysides should have some sort of purplish appearance. But what exactly will the color look like?

John Moores, a graduate student at the University of Arizona, and his advisor, Peter Smith, the PI of the Phoenix Mars Lander have built an optical setup which can generate (in an integrating sphere) the composite color that corresponds to any pre-specified optical spectrum. [Moores' weblog is here]. Their goal is to obtain a source that simulates the Martian lighting environment. Given the distinctly unearthly cast to the Martian illumination, it’s of interest to see whether long-duration exposure leads to adverse psychological or even physiological effects.

Moore used Fortney’s spectrum as an input to his apparatus to produce the image below. It shows the integrating sphere bathed in the resultant color. There is a small hole cut in the back of the sphere to allow access to a fiber spectrometer (visible as a dark spot to the left of center).

Dayside color of a hot Jupiter

To the best of our knowledge, therefore, the daysides of hot Jupiters are imbued with distinctly trendy distinctly mauve-like hues.

Categories: worlds Tags: