Controversy and Developments for the Gliese 581 System- Habitable or Non-Existent Planets?

by Dr. Tahir Yaqoob on November 29, 2012

The Gliese 581 (or GJ 581) system, and the number of planets it hosts, has gained considerable fame and notoriety. The system is centered around a dwarf star about a third of the mass of the Sun, located at about 20.3 light years from us. Fierce debate has prevailed in the literature, regarding the number of planets that GJ 581 hosts. The number ranges between 3 and 5 (at one time it was 6). No matter whether they actually exist or not, two of the potential planets, GJ 581 d and GJ 581 g were hailed in mid-2012 as amongst the top 5 potentially habitable planets outside our solar system (see for example, the top 5 potentially habitable planets list proposed by the Planetary Habitability Laboratory, or PHL). Considerable fanfare in the news everywhere accompanied this grand achievement of the objects labeled GJ 581 d and GJ 581 g which could either be entire worlds in an exotic star system, or artifacts of statistical noise. The prematureness of celebration rather takes away some glitz from the day when an irrefutable planet is actually found in a habitable zone. The controversy about the reality of some of the planets in the GJ 581 system of course continued in the literature after the celebrations were over. So what is the status now?

Several studies by different groups using some of the same data have come to different conclusions. The process of extracting information from the data is challenging: periodicities have to be identified by recasting the time-domain data into what is essentially a frequency histogram (or periodogram). One hopes to see peaks in such a periodogram corresponding to the orbital periods of planets. However, the procedure is riddled with difficulties because of imperfect sampling and noise. The name of the game then becomes that of determining if any peaks that you see above the surrounding noise are statistical artifacts. One method of doing this is to perform brute-force simulations with known starting configurations.

The two planets GJ 581 b and c are quite secure as they have been independently detected by two distinct instruments and seem to be robust against different noise models that are assumed. GJ 581 b has a minimum mass similar to Neptune but an orbital period of just over 5 days (so there is no counterpart to it in our solar system). The two planets GJ 581 c and d are supposed to be super-Earths with minimum masses of a few times that of Earth and orbital periods of about 13 and 67 days respectively. The candidate GJ 581 e is supposed to be very close-in to the host star, with an orbital period of just over 3 days and a minimum mass of only a couple of Earth masses. The potential planets GJ 581 f and g are the most controversial, the former would be far out in the system with an orbital period of 433 days, and the latter closer in at about 37 days. Minimum masses have been estimated at 7 and 3 Earth masses for GJ 581 f and GJ 581 g respectively, with the latter attaining fame for being a place where liquid water might exist.

A paper by R. Baluev that appeared in September 2012 went one step further and boldly stated that GJ 581 f and GJ 581 g are in fact artifacts (the word used was “illusions”) of noise. (The paper also has a good summary of the history of all the planet candidates and claims and counter-claims in the GJ 581 system.) The reality of GJ 581 d is left open, leaving only 3 firm planets in the system. The issue with the fallen planets GJ 581 f and GJ 581 g is so-called “red noise” as opposed to white noise. The latter is uniform with respect to frequency (reciprocal of orbital period), whereas the former has a skewed distribution in the periodogram, increasing, as you might expect as the period approaches the longest sampling time in the data. It seems that the exoplanet community is reinventing the wheel and rediscovering something that has been known for some time in other areas. For example, in the 1980s and 1990s there were many claims and counter-claims for finding periodic signals in the X-ray power output from systems associated with massive black holes at the centers of what are known as active galaxies. Today, none of the results from those two decades stands, and myself and some colleagues were involved in one of those “nails in the coffin,” showing that significance levels of periodicities had been overoptimistic by not rigorously modeling the red noise. (This sort of thing does not make you popular, in case you’re wondering.)

Still, people are reluctant to accept change. Only four days ago as I write, an article entitled Top Exoplanets for Alien Life: Photos appeared in Discovery News. Given that people regard the source, Discovery News, as authoritative and trustworthy, this is sorely disappointing and cringeworthy. Including the word photos in the title without any qualifier in the article that these are not images of real exoplanets is in itself an abandonment of responsibility. It’s the content that is even more disturbing. The images depict (yet another) top 5 habitable planet candidates list (not the same top 5 mentioned earlier). There are a series of images, with extended annotation, the first of which shows GJ 581 d. We know from the earlier discussion that the evidence for the existence of this planet is marginal, but we are assured by Discovery News that “If we ever need a new Earth and have the means to get there, Gliese 581 d may be our best bet for now.” Okay, next we have GJ 581 g. We know from the discussion earlier that this object does not exist. The Discovery News article does actually say that it “might not exist,” but then what is it doing in the top 5? What happened to other candidates that do actually exist and that were in the Planetary Habitability Laboratory list? Next, we have GJ 1214 b, which is okay, but bizarrely for the 4th out of the top 5 is the well-known hot Jupiter, HD 209458 b. This is one of the best-studied hot jupiters. It is extremely close to the host star, is a gas giant, and appears to be in disequilibrium due to large differential forces. It is not even in the habitable zone. Why it is in this top 5 list eludes me. Finally, number 5 is Kepler-10 b. The only reason cited is that this exoplanet has good evidence that it is rocky. That doesn’t really have anything to do with the prospects for alien life, one way or the other. It is not in a habitable zone either (although, even that may not of course be a necessary condition for life). I hope readers of this blog will enlighten those they come across who may be compromised by all of the distorted reporting that is out there.

Quickie General Updates

As of November 29, 2012, the number of confirmed exoplanets has not gone up by much since last time: 852 planets in 671 planetary systems. The number of confirmed planets observed by Kepler is now 105, more than 5 times greater than it was a year ago. Talking of Kepler, NASA has approved extension of the mission. The original length was three and a half years, and the extension could be as long as another four years (see news release).

Some of you may already know that the record holder for the closest exoplanet to Earth became alpha Centauri B b. The loser, eps Eridani b, had held on to the spot for over a decade. The distance between alpha Centauri B b and Earth is about 4.24 light years, less than half of the previous record. So the flight time to the nearest exoplanet has more than halved! Still, at 20,000 miles per hour, it would take 142,000 years or so to get there. There is no size estimate for the planet, and its minimum mass is about 1.1 Earth masses. The exoplanet alpha Centauri B b is a close-in planet, the orbit having a period of a mere 3.23 days. All this should be tempered by the fact that the existence of alpha Centauri B b is probabilistic (see the original Nature paper by Dumusque et al.).

Previous post:

Next post: