4 ms·
"After its discovery, it was generally believed that any atmosphere thick enough to keep the planet warm would become cold enough on the night side to freeze ou
by jerf 16y ago
"After its discovery, it was generally believed that any atmosphere thick enough to keep the planet warm would become cold enough on the night side to freeze out entirely, ruining any prospects for a habitable climate.... To test whether this intuition was correct, Wordsworth and colleagues developed a new kind of computer model capable of accurately simulating possible exoplanet climates.... To their surprise, they found that with a dense carbon dioxide atmosphere -- a likely scenario on such a large planet -- the climate of Gliese 581d is not only stable against collapse, but warm enough to have oceans, clouds and rainfall."
Question. How the bloody hell do you validate a computer model of how an atmosphere unlike anything we can observe in our solar system works? Did someone fly out there and verify it before publishing? And I'm supposed to believe your computer model iterates correctly over geologic time? Really? Really?
(Obligatory-but-totally-serious that this is still cool work not counting the computer model part, and to be honest I don't really care either way what the computer model says; I would be equally skeptical if they produced a model that claimed any outcome at all.)
- saulrh 16y agoSensitivity analysis. Elements of the simulation are based wholly on proven, low-level physics (for example, calculating the amount of heat absorbed by a given amount of CO2 from a given amount of sunlight) whenever possible, and when that's infeasible the modelers test a range of possible values to see what happens. For example, you can't exactly calculate the density of condensation nuclei in the atmosphere (used for calculating cloud formation behavior), so they just ran the simulation with a bunch of believable values. Another example is pretty clearly laid out in the first sentence of the "results" section: We performed simulations with 5, 10, 20 and 30 bar atmospheric pressure and 1:1, 1:2 and 1:10 orbit-rotation resonances for both rocky and ocean planets (see Table 1).
- jerf 16y agoThere is not enough data in a vague orbit and vague size (in a dataset that until recently contained a spurious planet) to feed a simulation like this, which means that the simulation was, mathematically necessarily (and I do not use this term lightly), made up from nearly whole cloth. And then we have no way to verify anything that it did say, assuming it even started from the correct place, especially with regards to stability on geologic time. Even an 99.9999999%-per-hour perfect simulation will still diverge on geologic time scales. For all we know, an atmosphere with a mountain range here will freeze out in a century, whereas a mountain range there could in fact result in a stable atmosphere over geologic time. This is all information-theoretic necessity. This is entertainment, not science. That may be all the scientists claimed. It's still a fun result, even if I wouldn't put my grandchildren on a flight to Gliese on the strength of it. I'm reacting against the certainty expressed in the article.
- saulrh 16y agoOk, I think I see what you mean now, and I agree with you. They've shown that the distribution assuming some set of priors includes some particular, interesting outcome. This is the first planet we've ever seen where we can say that given how much information we have about it, though, which means that it's still cool.
- robryan 16y agoYeah, we can still look at some extremely broad possibilities how how such a planet atmosphere could behave based on the little we know about the planet. Then the simulation gives some guide onto how each of these would play out. Sure it's very speculative, but one step up from no data at all, which is the best we have now. At least have an idea of the size, the orbit and the sunlight intensity that would be hitting the planet.
- fhars 16y agoBy making predictions that can be checked with the next generation of telescopes. That's how science works: look at what you have, figure out what you still want to know and design an experiment that can discriminate between different alternatives.
- mturmon 16y agoIt appears that radiative transfer modeling is a specialty of the LMD at CNRS in Paris, which is the home institution of the authors: http://en.wikipedia.org/wiki/GEISA http://en.wikipedia.org/wiki/GEISA The state of the art in radiative transfer modeling is surprisingly advanced. For more, see: http://en.wikipedia.org/wiki/Atmospheric_radiative_transfer_codes http://en.wikipedia.org/wiki/Atmospheric_radiative_transfer_... For instance, the use of RTMs for retrieving estimates of gas concentrations on Earth is routine. This idea is what the global CO2 maps produced by AIRS http://airs.jpl.nasa.gov/science/geophysical_science/ http://airs.jpl.nasa.gov/science/geophysical_science/ are based on. You can sense back-scattered radiation, which tells you about the CO2 concentration in the air below the satellite. A good RTM is what allows the inversion of radiation into gas concentration. Another related data point is that lots of investigators have been working for years on climate of Jupiter and Saturn using some of the same ideas. There are conference sessions about this topic, e.g. http://www.agu.org/meetings/sm05/sm05-sessions/sm05_SA24A.html http://www.agu.org/meetings/sm05/sm05-sessions/sm05_SA24A.ht... I'm not saying that this means the authors are right, just that there's more valid science here than you might guess.
- danssig 16y agoWelcome to science today. To their credit, at least when we find out the model was missing some key information they will update it and update their theories.
- jules 16y agoEspecially since researchers aren't able to accurately simulate Earth's climate over a smaller timescale, despite having much much better data about the current climate. Heck, for that planet they don't even know the composition of the atmosphere.