Possible solution to the riddle of HD 82943 multiplanet system: the three-planet resonance 1:2:5?

We carry out a new analysis of the published radial velocity data for the planet-hosting star HD82943. We include the recent Keck/HIRES measurements as well as the aged but much more numerous CORALIE data. We find that the CORALIE radial velocity measurements are polluted by a systematic annual variation which affected the robustness of many previous results. We show that after purging this variation, the residuals still contain a clear signature of an additional $\sim 1100$ days periodicity. The latter variation leaves significant hints in all three independent radial velocity subsets that we analysed: the CORALIE data, the Keck data acquired prior to a hardware upgrade, and the Keck data taken after the upgrade. We mainly treat this variation as a signature of a third planet in the system, although we cannot rule out other interpretations, such as long-term stellar activity. We find it easy to naturally obtain a stable three-planet radial-velocity fit close to the three-planet mean-motion resonance 1:2:5, with the two main planets (those in the 1:2 resonance) in an aligned apsidal corotation. The dynamical status of the third planet is still uncertain: it may reside in as well as slightly out of the 5:2 resonance. We obtain the value of $\sim 1075$ days for its orbital period and of $\sim 0.3 M_{\rm Jup}$ for its minimum mass, while the eccentric parameters are uncertain.