Skip to main content

Article Category


Article available in the folowing languages:

PASTIS helps confirm exoplanet existence with novel software

While the recent increase in the detection of exoplanets raises a distant hope of finding another habitable world, it also begs questions related to the accuracy of current detection software. Is there really an exoplanet where we thought there was one? The PASTIS project is coming up with new software to clear these doubts.

New exoplanet detection instruments with improved capabilities are starting to be rolled out. But as highly performing as these technologies may be, the detection and characterisation of extrasolar low-mass planets is currently limited by various astrophysical and instrumental effects. The precise characterisation of the host star, the modellisation of stellar jitter and the existence of false-positive scenarios are particularly challenging. According to Prof. Alexandre Santerne, Marie Curie fellow from the Centre for Astrophysics at the University of Porto in Portugal and coordinator of the PASTIS (Planet Analysis and Small Transit Investigation Software) project, overcoming these obstacles to Earth-like planet detection will require major improvements in analysis software so as to take these limitations into account. The two-year project aimed first at implementing such improvements into Prof. Santerne’s state-of-the-art PASTIS software, after which he hoped to validate new small and low-mass planets among the CoRoT, Kepler and HARPS data using the freshly-improved PASTIS software. What would you say are the main limitations of current data analysis software? In the field of exoplanets — at the very least — the main limitation we face is the current lack of knowledge of the host stars and in particular their variability (granulation, spots, magnetic cycles). Even for the Sun, recent high-precision observations showed that state-of-the-art analysis techniques would likely fail in finding or characterising inner solar system planets. Because of this lack of understanding of stellar variability, analysis software is not able to disentangle variability signals from planetary signals. This means that a substantial fraction of small planets reported so far might not be there. As I always say during outreach talk, if we find an Earth-like planet in the habitable zone of the nearest star, we might try to build a (very expensive) mission to visit this planet. But what will happen if in the end there is no planet? It would clearly be a disaster! The PASTIS software existed before the actual project started. How does it compare to other state-of-the-art solutions? The PASTIS software is conceptually completely different from other solutions in Europe. While most state-of-the-art software has been designed to derive the most precise and accurate parameters for a planetary system, PASTIS has a different objective: it aims to estimate the probability of a given planetary system being real and not mimicked by something else, in most cases whatever the parameters are. To me, this is a fundamental difference. We do not really care if a planet has an orbital period of 12.1 or 12.2 days, but we should care whether a planetary system is real or not. This is exactly what PASTIS is doing. What kind of software improvements did you aim for? The initial goal was to include stellar models into PASTIS, to better validate small exoplanets and characterise them, in particular in the case of stellar variability. We are now only a few months away from the end of the project. Would you say that you are on track to meet your initial objectives? Not completely. Some of the initial objectives were tested but resulted in a substantial slowing down of the software, to a level that would make the validation of new planets within the duration of the project impossible. Therefore, I searched for alternative solutions to meet the initial objectives, keeping the code relatively fast to run. I succeeded for the most part. The main objectives in terms of software improvement have now been implemented and validated. What was the nature of that problem? Well, the main difficulty was in improving the software in terms of complexity without increasing computing time. This was clearly the main challenge of this project and it required alternative solutions. As always in physics, when we are faced with an overly complex system, we try to find reasonable approximations. What about the validation of new planets, which was also one of your main objectives? New planets were indeed validated using PASTIS and published: K2-19 b & c (a system of two Neptune-size planets — Armstrong, Santerne et al., 2015) and WASP-121 b (an interesting hot Jupiter — Delrez, Santerne et al., 2016). More systems are currently being analysed, including Earth-size planets. What are your plans until and after the end of the project? Thanks to this Marie Curie fellowship, I just got a faculty position in France / Marseille, starting this fall. So, I will continue developing and using PASTIS to validate exoplanets and for the preparation of the ESA PLATO space mission. In particular, I plan to exploit the new data from the GAIA mission to improve the validation of planets with PASTIS. The improvement of the code, making it even faster, is also among my top priorities for the next few years. PASTIS Funded under FP7-PEOPLE project page on CORDIS