Lessons learned from the detection of wide companions by radial velocity and astrometry

ABSTRACT The detection and constraint of the orbits of long-period giant planets is essential for enabling their further study through direct imaging. Recently, it has been highlighted that there are discrepancies between different orbital fitting solutions. We address these concerns by re-analysing the data for HD 28185, GJ 229, HD 62364, HD 38529, 14 Her, $\epsilon$ Ind A, HD 211847, HD 111031, and GJ 680, offering explanations for these discrepancies. Based on the comparison between our direct modelling of the astrometric catalogue data and the orvara code, we find the discrepancies are primarily data-related rather than methodology-related. Our re-analysis of HD 28185 highlights many of the data-related issues and particularly the importance of parallax modelling for year-long companions. The case of eps Ind A b is instructive to emphasize the value of an extended radial velocity baseline for accurately determining orbits of long-period companions. Our orbital solutions highlight other causes for discrepancies between solutions including the combination of absolute and relative astrometry, clear definitions of conventions, and efficient posterior sampling for the detection of wide-orbit giant planets.