i wear my sunglasses at night

Of all the photographs that our robot emissaries have radioed back to Earth, my vote for the most stunning is the Hubble ACS image of the “Sombrero Galaxy”, M104. The glow of its halo makes the the idea of 100 billion stars seem comprehensible.

HST ACS mosaic of M104

It’s important to remember, however, that the Hubble image is actually a long CCD time-exposure to light gathered by a 240 cm mirror. If you could be somehow transported to a location in space where M104 looms large in the sky, you would see that HST imparts a severely inflated expectation. From a distance, say, of 300,000 light years, M104 would be so dim that you would see only a faintly ominous, faintly glowing flying saucer.

Approximate naked-eye view of M104 from ~300,000 light years distance

Indeed, the great Andromeda Galaxy, M31, subtends an angle larger than the full Moon in the sky, and it is literally almost directly overhead right now (9:36 PM, Dec 3, latitude 36.97 deg N). The storms from earlier this week have blown through. The sky sparkles with brilliant clarity. Yet when I step outside and look up, I can’t see the Andromeda Galaxy at all. It’s too faint. In a 1:10,000,000,000,000 scale model of M31, the stars are like fine grains of sand separated by miles. Our Galaxy, the Andromeda Galaxy, and the Sombrero Galaxy are all essentially just empty space. To zeroth, to first, to second approximation, a galaxy is nothing at all.

A Hot Jupiter, on the other hand, seen at similar angular size, is undeniably impressive.

HD 149026 b in the crescent phase

The dayside, blindingly illuminated by the scorching proximity of the star, is roughly 500 times brighter than desert sand dunes on a midsummer day. In order to look at the illuminated side of the planet at all, you need extremely dark wraparound sunglasses, or better yet, an eyeshield made from #10 welders glass (where #14 welder’s glass is recommended for those who stare at the sun).

With the brilliance of the dayside cut to a manageable level, what would you see? The majority of the light coming from the planet is simply reflected starlight. If the planet uniformly reflects the light that strikes it, then you simply see a blank white surface if the parent star is similar to the Sun, and a yellow-orange to orange-red expanse if the parent star is a cooler K-type or M-type dwarf star.

The gases that make up the outer layers of the planet do not reflect all frequencies of light equally, however. The air of the outer layers of a hot Jupiter is a scaldingly toxic witches brew of hydrogen, helium, steam, methane, ammonia, cyanide, acetylene, hydrogen sulfide, soot, and a whole host of other hardy, reactive, and generally unpleasant compounds.

In our solar system, for example, Uranus and Neptune have distinctive blue-green casts because at the level in their atmospheres where light is primarily reflected, the ambient methane gas is highly effective at absorbing red frequencies. The originally white sunlight is reflected with a blue-green hue by the selective removal of red.

Uranus and Neptune (from Voyager II)

The photo (mosaic) below was obtained by the Cassini spacecraft as it was flung past Jupiter on its way to Saturn. The images were processed to give the same view that the naked eye would see. Jupiter reflects an enormous amount of detail from its cloudy face.

True-color Cassini mosaic of Jupiter

Across the swathes of Jupiter where the visible clouds tower to great heights, the eye sees regions that are frigid, eighty degrees colder than the depths of an Antarctic winter (-200 F). In such a cold environment, icy compounds of Ammonia are stable, and their presence lends the clouds a reddish hue. Jupiter’s Great Red Spot is an example of just such a topographic high.

On other regions of Jupiter’s visible surface, the atmosphere is transparent to greater depths. As on Earth, where clear skies are associated with dry air, so too on Jupiter. When we look down into the drier Jovian regions, we see to lower lying decks of cloud where the temperature is about the same as a chilly Arctic night. Here, the chemistry in the clouds causes their color to tend toward lighter shades, whites, beiges, ochers.

Like any non-transparent object, Jupiter glows with its own radiation. Because the outer layers of Jupiter are so cold, this intrinsic light lies in the infrared. Seen with an infrared detector (such as this view made at 5 microns with the NASA IRTF) Jupiter is a dramatic sight.

IRTF 5 micron image of Jupiter

In the rattlesnakes-eye view, the Red Spot forms an oval of relative darkness. The high clouds act like a blanket that blocks the warmer underlying layers from view. In the infrared, the dry areas, where we see the deepest, glow the brightest. In an ironic twist of fate, the Galileo atmospheric probe parachuted into one of the driest regions of the Jovian atmosphere, a so-called 5 micron hot spot (circled in the image above).

On a hot Jupiter, the surface gas is heated to temperatures in the 1000-1500 K range on the dayside. Computer simulations show that winds of hellacious strength tear continually around the planet, carrying heat from the dayside and disgorging it into the night. The atmosphere on nightside glows brilliantly. Turbulent brick-red whorls merge into fiery tendrils of orange braided with dazzling white.

A most eccentric character

Of all the known extrasolar planets, HD 80606b — in both the technical and the colloquial sense — is the most eccentric. This world has at least five times the mass of Jupiter, and it circles its parent star on an extremely elongated 111.4 day orbit:

planetary orbit for HD 80606 b

Today (as seen from Earth!) HD 80606b is still near the far point of its orbit, at a distance of about 0.85 AU from the central star. The temperature in the upper atmospheric layers of the night-side has possibly dipped low enough so that torrential rains and violent thunderstorms are rumbling across its vast billowing horizons. During the rest of December and through most of January, the planet will fall in almost the full distance to the star, eventually swooping within 6 stellar radii as it whips through periastron. On January 26th, at the moment of closest approach, the temperature at the cloud tops will exceed 1000 Kelvin. The auroral displays will be dramatic beyond compare, and indeed, during the days to either side of periastron passage, it might be worth tuning in to the planet on the decameter band.

The discovery of the planet and its orbital solution were announced by the Geneva Observatory Planet Search Team in an April 04, 2001 ESO press release, and the radial velocities have since been made publicly available (right on!) at the CDS repository (see Naef et al 2001). You can therefore use the systemic console to fit this system and examine how radial velocity curves behave for extremely eccentric orbits.

The star HD 80606 is accompanied by a visual binary companion, HD 80607. The projected separation of the two stars is 2000 AU (fifty times the Sun-Pluto distance). When the HD 80606 b travels through the segment of its orbit that lies between the two stars, the night-side cloud tops of the planet are lit by the distant binary companion to ambient brightness that is very similar to a fully moonlit night on Earth. The two stars have similar masses, sizes, and temperatures to the Sun, but, like many of hosts of short-period massive planets, they are enriched in “metals” (gold, chromium, iron, carbon, oxygen, etc. etc.) by a factor of more than two relative to the solar value.

How did the planet get into its weird orbit?

Wu and Murray (2003) have suggested that HD 80606b’s extreme eccentricity is the result of a three-body interaction known as the “Kozai effect” between the planet and the two stars.

The next big thing

We know that planets aren’t rare, and by now, with the tally over at the extrasolar planet encyclopedia poised to blast past 200, the announcement of a newly discovered run-of-the-mill Jupiter-sized planet barely raises the collective eyebrow.

The headline that everyone is anticipating is the discovery, or better yet, the characterization of a truly habitable world — a wet, Earth-sized terrestrial planet orbiting in the habitable zone of a nearby star. Who is going to get to this news first, and when?

299 million dollars of smart money says that Kepler, a NASA-funded Discovery mission currently scheduled for launch in June 2008, will take the honors. The Kepler spacecraft will fly in an Earth-trailing 377.5 day orbit, and will employ a 1-meter telescope to stare continuously (for at least four years straight) at a patchwork of 21 five-square-degree fields of the Milky Way in the direction of the constellation Cygnus. Every 15 minutes, the spacecraft will produce integrated photometric brightness measurements for ~100,000 stars, and for most of these stars, the photometric accuracy will be better than one part in 10,000. These specs should allow Kepler to detect transits of Earth-sized planets in front of Solar-type stars.

Kepler has a dedicated team, a solid strategy, and more than a decade of development work completed. It’s definitely going to be tough to cut ahead of Bill Borucki in line. Does anyone else stand a chance?

Practitioners of the microlensing technique have a reasonably good shot at detecting an Earth-mass planet before Kepler, but microlensing-detected planets are maddeningly ephemeral. There are no satisfying possibilities for follow-up and characterization. Doppler RV has been making tremendous progress in detecting ever-lower mass planets, but it seems a stretch that (even with sub-1 meter per second precision) the RV teams will uncover a truly habitable world prior to Kepler, although they may well detect a hot Earth-mass planet.

There is one possibility, however, whereby just about anyone could detect a habitable planet (1) from the ground, (2) within a year, and (3) on the cheap. Stay tuned…

Now fielding three tutorials

Three detailed console tutorials have recently been developed, and are now online at oklo.org.

Tutorial #1 steps through the basic features of the console, using the published radial velocity data-set for the Jupiter-like planet orbiting HD 4208.


Tutorial #2
takes a more detailed look at the console, and shows how to use periodograms and multiple-planet fitting to recover the three planetary companions (the so-called Fourpiter, Twopiter, and Dinky) orbiting Upsilon Andromedae.


Tutorial #3
tackles the tough problem of multiple-planet fitting in the presence of planet-planet interactions, and uses the console to explore the remarkable, recently published Gl 876 data set.

The console has landed.

After more than a year of development work, the beta version of the systemic console java applet is now up and working at oklo.org. Hats off to Aaron Wolf for coding it into reality.

In a series of posts, we will look in detail at the organization, operation, and features contained in the console. For now, however, rev up your G4s and your G5s, take it for a spin, and let us know how it works for you.

The current location for the console is:

www.oklo.org/SystemicBeta/SystemicBeta.html.

It’s also accesible from the menu bar to the right. At the moment it has been tested only with Safari 2.0.2 running on OSX 10.4.3. Firefox 1.0.6 still seems to have issues with the applet. We’ll resolve these first, and then (with CDR Paul Shankland leading the charge) we’ll move on to thwart Bill Gates’ best attempts to protect the MS Explorer user base from Systemic’s seductive charms…

All hands on deck (GJ 876)

nsf illustration of GJ 876 d

Paul Shankland has been visiting Santa Cruz this week, and everyone agrees that it’s about time to get the GJ 876 transit situation sewed up once and for all. Aquarius is still up in the early evening, and a planet “c” transit opportunity is bearing down with 30.1 day semi-clockwork precision. So out went the following alert to the transitsearch.org e-mail list:

Thursday Afternoon, Nov. 17, 2005

Dear Transitsearch Observers,

We’d like to alert you to an opportunity to check the GJ 876 system for planetary transits. Photometry is desired during a twelve-hour window centered on JD2453693.491 (Friday Nov. 18, 23:47 UT).

As you have likely heard, the GJ 876 system was recently found to harbor a low-mass (7.5 Earth Mass) planet on a 1.94 day orbit. The new planet is referred to (rather prosaically) as GJ 876 “d”, and is the third planet detected in the GJ 876 system. The discovery paper is scheduled for an upcoming issue of the Astrophysical Journal, and is also available on the astro-ph preprint server: http://arxiv.org/abs/astro-ph/0510508

Sadly, transits for planet “d” have been ruled out to high confidence.

As a result, however, of (1) inclusion of the third planet in the dynamical model for the system, and (2) a large number of new high-precision radial velocities, Eugenio Rivera has produced new transit ephemeris predictions for the outer two planets in the GJ 876 system. These differ by several hours from the dynamical predictions that are currently posted on the transitsearch.org candidates site, e.g.:

http://www.ucolick.org/~laugh/GJ876____c.transits.txt

We’re working through an extensive analysis which shows that neither “b” nor “c” is transiting, but this analysis is nevertheless in great need of observational verification. There is a conflict between dynamical fits to the radial velocities (which indicate that the system is inclined by 50 degrees to the plane of the sky) and the results of Benedict et al (2002, ApJL 581, 115), who used HST to get astrometric measurements that suggest a nearly edge-on configuration.

We’d thus like to request photometry of the star to six hours on either side of JD2453693.491 (Friday Nov. 18, 23:47 UT).

Information regarding observing GJ 876 and photometry submission instructions are at:

http://www.aavso.org/news/ilaqr.shtml

Additional background is on the transitsearch GJ 876 results page:

http://www.ucolick.org/%7elaugh/GJ876____c.results.html

Note that this page states that the photometric campaign is over, but the new dynamical model indicates that more photometry is desirable.

Other observing opportunities are (were) as follows:

For planet c (the middle one):

predicted central transit (UT)
————————————
2005 Aug. 20 15:40
2005 Sep. 19 18:41
2005 Oct. 19 20:53
2005 Nov. 18 23:47
2005 Dec. 19 01:36

For planet b (the outer one):
————————–
2005 Aug. 22 17:28
2005 Oct. 22 17:34
2005 Dec. 22 18:05

Finally, we’d like to thank everyone for being patient over the 8 months, during which we have not been running coordinated campaigns. With Shankland of USNO “on the bridge”, we’re now ramping up for a more active phase. Stay tuned!

The music of the spheres (sounds terrible)

After using the console for a while, you’ll notice that it’s often easy to find a reasonably good (say, chi-square of 3-5) multiple-planet fit to a given radial velocity data set. This rule of thumb tends to be especially true if you allow the planets to have large eccentricities. But how does one know whether the fit is likely to be correct?

This is one of the questions that the systemic simulation is designed to answer.

Most of the time, however, if a fit contains large enough eccentricities for the planet orbits to cross, then the trial system will be dynamically unstable. That is, the planets in the model will suffer a close encounter, which is generally followed (or directly accompanied) by a disaster. The planets collide, or one or more of them is ejected, or one of them is thrown into the central star.

While it is certainly true that such catastrophes have been reasonably common throughout galactic history, it is exceedingly unlikely that any particular planetary system that we observe will be on the verge of a dramatic instability. The stars that can be observed using the Doppler radial velocity method are billions of years old. If a star had an unstable planetary system, it is likely that the instability either occurred long ago, or that won’t happen for a long time to come.

As a result, an important requirement for any radial velocity fit is that it correspond to a dynamically stable system. Traditionally, this can be checked either by integrating the system forward in time, or by applying a technique which checks for the presence of chaos in the orbits. (Indeed, all of the planetary systems that underlie the systemic database have been integrated for one million orbits prior to being “observed”. These pre-integrations establish a strong likelihood of short-term dynamical stability for all the systemic systems.)

Here’s an idea that sounds possibly promising. If the radial velocity waveform of a planetary system is converted into an audio signal, is it possible for the human ear to rapidly detect whether a system is likely to be unstable? To test this, we’re working on bringing an audio generator into the systemic console.

More generally, what do the extrasolar planetary radial velocity reflex waveforms sound like? The short answer is, they sound terrible. There are interesting reasons for this, which we’ll pick up in a future post. For now, have a listen to these .wav’s (created by Aaron Wolf) of two of the best-known multiple planet systems: GJ 876 and Upsilon Andromedae

And try to listen for the (heavily processed voice of the better-voice-of-the-two GJ876 in the forthcoming James Alley Remix).

Hello world.

What is systemic?

Systemic is a public research collaboration. Systemic’s goal is to obtain a better understanding of the census of planets in the galaxy.

The systemic blog, hosted by oklo.org, provides a framework for updates and information relating to the collaboration. It also serves as an online forum for discussion of extrasolar planets.