Objectives
The objective of this service is to provide to the user orbital parameters fitted with a MCMC code (described below) from data provided by the user, first based on a on-request protocol, and later via a dedicated interface.
Current status
The DIVA+ database is about to offer within a few months an on-line orbital fitting service for exoplanets and/or (sub)stellar companions based on astrometric high contrast imaging and/or radial velocity data. The data maybe archival data available in our DIVA+ data base, possibly completed by private data supplied by the user.
The service is not yet available in an automatic way, but if you are interested by this service, please contact Hervé Beust (herve.beust at univ-grenoble-alpes.fr). Do not hesitate to send us comments (needs, ideas, …) on this service.
Methodology
This facility will operate a home-developed MCMC code, which will keep being developed in upcoming years. This constitutes an alternative to the public orbitize! code.
Already available options include:
- A Python exploitation and visualization software
- A least-square Levenberg-Marquart fitter, that can be used as a quick alternative to MCMC for a rapid view, or prior to the launch of a longer MCMC run. The least-square facility runs within a few seconds and can helps to derive a first overview of the seeked orbit(s).
- The possibility to fit several orbits and masses simultaneously in a given multiple system, based on a Jacobi coordinates formalism.
- The possibility to introduce additional priors, especially on individual masses, and moreover linear combinations of those masses. This allows for instance to add a prior on a dynamical mass involving the sum of several masses.
- A Python exploitation and visualization software
- A least-square Levenberg-Marquart fitter, that can be used as a quick alternative to MCMC for a rapid view, or prior to the launch of a longer MCMC run. The least-square facility runs within a few seconds and can helps to derive a first overview of the seeked orbit(s).
- The possibility to fit several orbits and masses simultaneously in a given multiple system, based on a Jacobi coordinates formalism.
- The possibility to introduce additional priors, especially on individual masses, and moreover linear combinations of those masses. This allows for instance to add a prior on a dynamical mass invoving the sum of several masses.
Future developments are:
- the user will eventually be able to operate on-line via a dedicated interface.
- A version of the code making use of universal Keplerian variables instead of classical variables will be available. This results in an extra cost in terms of computing complexity, but turns out to speed up convergence in systems with very eccentric orbits (Beust et al. 2016, A&A 587, A89).
Data needed to perform the orbital fits
- Time series of astrometric positions and/or radial velocities: Time series of relative astrometric position and stellar radial velocity data (as of yet). Future scheduled evolution of the service will include up to 4 types of data : relative astrometric positions, absolute astrometric data (Gaia..), stellar radial velocity data, and radial velocity data of imaged exoplanets relative to their host stars.
- Associated to uncertainties
Outputs
A standard output consists of a collection of 400,000 orbital solutions representative of the posterior distribution. Vizualization software is provided to exploit this file and generate various kinds of output, (corner plots, histograms, etc..). 2 examples are shown below.
