systemic

spectra1
greg posted in Uncategorized on February 22nd, 2007

Image Source.

So, uhh, yeah, the oklo blog went through a dry spell with no posts last week. This was primarily a consequence of the fact that Spitzer GO-4 proposals were due last Friday. I teamed up with Drake Deming of GSFC and UCSC physics grad student Jonathan Langton to propose a 30-hour observation of HD 80606b during the ‘606 day that’ll occur next November 20th. In an upcoming article, I’ll be pushing the reasons why we’re really excited about the possibility of observing HD 80606 b during its big periastron swing.

The Spitzer Space Telescope has turned out to be a regular wellspring of exo-planet results. It’s providing very interesting and often surprising constraints on the weather conditions at the surfaces of the hot Jupiters, and another big new result was announced today. Three different teams released the first-ever observations of emergent infrared spectra from two observational runs on HD 189733 and HD 209458.

The transiting planet HD 189733 b was discovered by the Swiss team in 2005. Of the fourteen known transiting planets, HD 189733 b is the best-suited for detailed follow-up observations. The parent star lies only 19 parsecs away, the orbital period is a skimpy 2.1 days, and the 1.15 Jupiter-mass planet has a radius fully 15% the size of the primary star’s radius. Like the other transiting systems, the planet, the orbit, and the star can all be drawn completely to scale on a “saved for web” diagram that’s only 420 pixels across:

Grillmair et al.’s Spitzer spectrum of HD 189733 was obtained with 12 hours of observation, in which the brightness of the star at infrared wavelengths between 7 and 14 microns is compared in and out of the secondary transit:

The observed flux distribution from the planet is nearly completely flat as a function of wavelength! Models of the atmospheres of hot Jupiters had all predicted that the presence of water vapor in the planetary atmosphere would lead to a prominent absorption feature at ~8 microns. No hint of the predicted dip was seen. The overall amount of infrared light coming from the planet during the secondary transit indicates that heat is probably being efficiently redistributed between the day and night sides of the planet.

A Nature paper by Jeremy Richardson and collaborators reported a very similar result for HD 209458 b. Their spectrum runs between 7.5 and 13.2 microns, is similarly devoid of absorption features, and also suggests a modest day-night temperature difference.

So how to interpret these results? One possibility is that the lack of absorption lines is caused by a high, uniformly emitting cloud layer, perhaps made of silicate grains. A problem with this interpretation, however, is that the cloud decks would have to be extremely dark and unreflective in the optical. Hot Jupiters absorb nearly 95% of the radiation that they receive from their parent stars. Another possibility, put forward by Jonathan Fortney, is that the atmospheres of these planets are isothermal down to large optical depths. Because we can’t actually see to a hotter underlying layer, there’s no mechanism for deep absorption lines to form.

And finally, another, rather startling, interpretation of the results was offered several hours ago by CBS News:

Image Source.

The exoplanet prediction market1
greg posted in systemic faq, exoplanet detection, worlds, non-technical posts on February 20th, 2007

Image Source.

At first glance, the market capitalization of the Chicago Board Options Exchange, and the list of astronomers active in the field of extrasolar planets would appear to have nothing to do with one another. These two disparate entities are connected, however, by the fact that they’ve both undergone explosive growth over the past decade, and both are continuing to grow. They signify highly significant societal trends.

I think it’s safe to predict that in 25 years, the market for financial derivatives, and the level of economic activity associated with exoplanets will both be far larger than they are now. It’s interesting to ask, will there be an unanticipated co-mingling between the two? And if so, how will it occur?

One very realistic possibility is the development of an exoplanet prediction market, in which securities are issued based on particular fundamental questions involving the distribution of planets in the galaxy. Imagine, for example, that you’re an astronomer planning to devote a large chunk of your career to an all-or-nothing attempt to characterize the terrestrial planet system orbiting Alpha Centauri B. In the presence of a liquid, well-regulated exoplanet prediction market, you could literally (and figuratively) hedge your investment of effort by taking out a short position on a security that pays out on demonstration of an Earth-mass planet orbiting any of the three stars in Alpha Centauri.

Prediction markets have been adopted in a very wide range of contexts, ranging from opening weekend grosses for big-budget movies, to forecasts of printer sales, to the results of presidential elections. A highly readable overview of these markets by Justin Wolfers (who was featured last week in the New York Times) and Eric Zitzewitz of the University of Pennsylvania is available here as a .pdf file. The ideosphere site contains a wide variety of markets (trading in synthetic currency) and includes securities directly relevant big-picture questions in physics, astronomy and space exploration. Here’s the price chart for the Xlif claim,

which pays out a lump-sum of 100 currency units if the following claim is found to be true:

Evidence of Extraterrestrial Life, fossils, or remains will be found by 12/31/2050. Dead or extinct extraterrestrial Life counts, but contamination by human spacecraft doesn’t count. (Life engineered or created by humans doesn’t count.) The Life must have been at least 10,000 miles from the surface of the Earth. If Earth bacteria have somehow got to another planet and thrived, it counts, as long as the transportation wasn’t by human space activities.

It’s very interesting to compare the bullish current Xlif price quote of 72 with the far more bearish sentiment on Xlif2, which is currently trading at an all-time low of 17,

and which pays out if “extraterrestrial intelligent life is found prior to 2050″, and more specifically,

Terrestrial-origin entities (e.g. colonists, biological constructs, computational constructs) whose predecessors left earth after 1900 do not satisfy this claim. If the intelligence of the ET is not obvious, the primary judging criteria will be either a significant level of technological sophistication (e.g. radio transmitting capability) or conceptual abstraction (e.g. basic mathematical ability). Radio signals received or similar tell-tale signs of intelligence (e.g. archeological discoveries) detected and accepted by scientific consensus as originating from intelligent extraterrestrials would satisfy the claim even if not completely understood by the claim judging date.

Recently, open-source software has been released that makes it straightforward to set up a prediction market. We’ll soon have the world’s first exoplanet stock market up and running right here at oklo.org. In the meantime, feel free to submit specific claims (in the comments section for this post) that might lend themselves to securitization…

glow0
greg posted in worlds, non-technical posts on February 12th, 2007

Image Source.

Saturn reached opposition yesterday, marking the moment in our yearly orbit when the Earth draws closest to the massive ringed giant. At midnight, Saturn is currently the only planet visible in the sky. It’s an odd feeling to stare at the bright unresolved spot of light that encompasses the planet, the rings, and the moons into a tiny golden point, and to know that Cassini, our robot emissary, is actually out there, almost a billion miles away, taking photograph after photograph, and radioing them back to a mere mouse click away.

Schematic image of the solar system on 2/11/2007 created at Solar System Live.

Saturn and its rings are good reflectors of light, but nevertheless, in the vicinity of the planet, the glare is far from overwhelming. The ambient light levels are only a bit more than 1% that of a bright summer day on Earth. It would be easy to stare at the crisply defined terminator marking the day-night boundary on the planet and the arcs of black shadow cast by the rings. On the Cassini website, there are many views that show the planet as it would appear to human eyes.

Image Source.

Cassini also has the ability to photograph in the infrared. The following false-color photograph shows visible and infrared images of the planet superimposed. In the infrared, Saturn glows with interior heat — still welling up from the planet’s formation — that illuminates the night from within.

Image Source.

The picture above is not a bad approximation of what a younger more massive planet would look like to the naked eye. 2M1207 b, for example, which seems to have a mass about five times that of Jupiter, is in a 1700-year orbit around a young 25-Jupiter mass brown dwarf. At 1250 Kelvin, 2M1207b is still warm enough to be self-luminous in the visible region of the spectrum. It is also slightly illuminated by the light of its companion (whose ~2500K surface is intrinsically 100 times more luminous.) Methane absorption and Rayleigh scattering of incident light in 2M1207 b’s atmosphere likely give the weak crescent a bluish-green hue.

Image Source.

Lonely Planet Guide to the Hyades13
greg posted in systemic faq, exoplanet detection on February 9th, 2007

Image Source.

It’s been a hectic week, and now that it’s February, my New Year’s resolution to write 2-3 posts per week managed to lose its shaky option on my priorities.

Eugenio stopped by my office this afternoon to outline his latest code developments for the console. He’s mostly finished implementing a Bulirsch-Stoer integrator. Once this algorithm is tested and operational, it will produce very significant speed-ups for the fitting and the stability analysis of tough multiple-planet systems such as 55 Cancri and GJ 876. Then it’ll be on to a rollout of the bootstrap method for computing uncertainties for the orbital elements in the planetary fits.

“So did you see the new planet?” he asked.

“Huh?” I hadn’t heard anything about it.

Turns out that Bunei Sato and his collaborators have detected a periodic radial velocity variation for the star Epsilon Tauri. Their preprint is on the Astrophysical Journal’s website, but it doesn’t seem to have hit the preprint server yet. This star is a prominent member of the nearby Hyades cluster, and is easily visible to the naked eye as part of the well-known “V”-shaped asterism near Aldeberan in the sky.

Image Source.

Eps Tau is bright enough to have garnered 40 different names in the Simbad catalog, and it’s now listed in the console menu and on the systemic backend as HD 28305. This is one of the most straightforward radial velocity datasets that you’ll come across, and thus makes a good system for first-time users to fit. A few debonair moves with the downloadable console conjure up a model planet with a period of 594 days, an orbital eccentricity e=0.15, and a minimum mass 7.6 times that of Jupiter:

Epsilon Tauri is one of the four stars in the Hyades that are currently nearing the end of their lives and are evolving through the red giant phase. It’s 14 times larger than the Sun, and it’s luminosity is 97 times the solar value. It weighs in at 2.7 solar masses, making it the most massive star known to harbor a planet.

So what’s the story? The Hyades are a metal-rich cluster. One would naively expect that the supersolar composition of the precursor star-forming giant molecular cloud would have lead to a large fraction of the cluster members harboring readily detectable planets. It’s also true that stars somewhat more massive than the Sun should harbor a higher-than-average fraction of giant planets. Eps Tauri scores on both counts.

[Note: John Johnson’s thesis work at UC Berkeley and Bunei Sato’s RV survey are both capable of providing observational support for the hypothesis of a positive correlation between the detectable presence of a planet and the mass of the parent star. See talk #1 on the Systemic Resources page for more details.]

Young Cluster NGC 3603, Source: NASA

It’s important to keep in mind, however, that a cluster environment will have a strong effect on giant planet formation. Currently, the Hyades are 600 million years old, and the cluster has lost a large fraction of its O.G.s to the general galactic field through the process of dynamical escape. If we extrapolate back to the cluster’s early days, we find that the Hyades would have resembled the Pleiades 500 million years ago, and would have looked like the Orion Nebular Cluster during the first few million years of its existence.

The UV radiation environment in the original Hyades cluster was fierce. The protostellar disks of the individual Hyads were likely photoevaporated before the growing planetary cores were able to reach the runaway gas accretion phase that gives rise to Jupiter-mass planets (see our paper on this topic). When we get the full inventory of planets in the Hyades, I think we’ll find plenty of Neptunes and terrestrial planets, but almost nothing in the Jovian range. Indeed, work by Bill Cochran and the Texas RV group has demonstrated that the Hyades are generally deficient in massive planets.

My guess is that Epsilon Tauri b is an example of a planet that formed through the gravitational instability mechanism. Gravitational instability should generally produce more massive planets (e.g. HIP 75458 b, and HD 168443 b and c) and its efficacy will be little-affected by UV radiation from neighboring stars. It likely occurs once per every several hundred stars that are formed, and so it’s perfectly reasonable that there’s one star in the Hyades that has a planet formed via the GI mechanism.

For more information, this series: 1, 2, 3, 4, 5, 6, and 7
of oklo posts compares and contrasts the gravitational instability and core accretion theories for giant planet formation.

a bunch of cool new stuff0
greg posted in systemic faq on February 2nd, 2007

Image Source.

Stefano and Eugenio have both been working hard on the systemic console and backend, and as a result of their efforts, we’re now able to roll out a number of new features.

The backend now features a systemic wiki in which users can collaborate on a wide variety of writing projects related to systemic in particular and extrasolar planets in general. Features include discussion pages for individual systems, the framework for a comprehensive console and backend manual, and an exoplanetary news wire. Our first news service is being provided by Mike Valdez, who combs astro-ph every day and extracts any new preprints that are germane to the those interested in exoplanets. Stefano wrote the code from scratch, so there are endless possibilites for customization. Give it a try.

On the console front, Eugenio has aggregated a uniform listing of the literature sources of all of the radial velocity data sets provided by the console. This information is in a file vels_list.txt, which is now included in the systemic.zip package. If you are using the console for scientific research that you intend to publish, it’s now a snap to get the correct citations for any of the individual systems included on the console.

Many users have expressed interest in what our own solar system would look like to a dedicated radial velocity observer on another star. Eugenio has put together an expansion pack that contains 17 manufactured data sets based on the Solar System. A second expansion pack contains an analogous set of manufactured data sets for various plausible configurations of planets orbiting Alpha Centauri A and B. Both are available on the downloads page for the downloadable systemic console.

Check it out!