Keywords

Gliese 876 harbors one of the most dynamically rich and well-studied exoplanetary systems. The nearby M4V dwarf hosts four known planets, the outer three of which are trapped in a Laplace mean-motion resonance. A thorough characterization of the complex resonant perturbations exhibited by the orbiting planets, and the chaotic dynamics therein, is key to a complete picture of the system's formation and evolutionary history. Here we present a reanalysis of the system using 6 yr of new radial velocity (RV) data from four instruments. These new data augment and more than double the size of the decades-long collection of existing velocity measurements. We provide updated estimates of the system parameters by employing a computationally efficient Wisdom–Holman N-body symplectic integrator, coupled with a Gaussian process (GP) regression model to account for correlated stellar noise. Experiments with synthetic RV data show that the dynamical characterization of the system can differ depending on whether a white-noise or correlated-noise model is adopted. Despite there being a region of stability for an additional planet in the resonant chain, we find no evidence for one. Our new parameter estimates place the system even deeper into resonance than previously thought and suggest that the system might be in a low-energy, quasi-regular double apsidal corotation resonance. This result and others will be used in a subsequent study on the primordial migration processes responsible for the formation of the resonant chain.

Continued radial velocity (RV) monitoring of the nearby M4V red dwarf star GJ 876 with Keck/High Resolution Echelle Spectrograph has revealed the presence of a Uranus-mass fourth planetary companion in the system. The new planet has a mean period of P_e = 126.6 days (over the 12.6-year baseline of the RV observations), and a minimum mass of m_esin i_e = 12.9 ± 1.7 M_⊕. The detection of the new planet has been enabled by significant improvements to our RV data set for GJ 876. The data have been augmented by 36 new high-precision measurements taken over the past five years. In addition, the precision of all of the Doppler measurements have been significantly improved by the incorporation of a high signal-to-noise template spectrum for GJ 876 into the analysis pipeline. Implementation of the new template spectrum improves the internal rms errors for the velocity measurements taken during 1998–2005 from 4.1 m s⁻¹ to 2.5 m s⁻¹. Self-consistent, N-body fits to the RV data set show that the four-planet system has an invariable plane with an inclination relative to the plane of the sky of i = 595. The fit is not significantly improved by the introduction of a mutual inclination between the planets "b" and "c," but the new data do confirm a non-zero eccentricity, e_d = 0.207 ± 0.055 for the innermost planet, "d." In our best-fit coplanar model, the mass of the new component is m_e = 14.6 ± 1.7 M_⊕. Our best-fitting model places the new planet in a three-body resonance with the previously known giant planets (which have mean periods of P_c = 30.4 and P_b = 61.1 days). The critical argument, φ_Laplace = λ_c − 3λ_b + 2λ_e, for the Laplace resonance librates with an amplitude of Δφ_Laplace = 40° ± 13° about φ_Laplace = 0°. Numerical integration indicates that the four-planet system is stable for at least a billion years (at least for the coplanar cases). This resonant configuration of three giant planets orbiting an M dwarf primary differs from the well-known Laplace configuration of the three inner Galilean satellites of Jupiter, which are executing very small librations about φ_Laplace = 180° and which never experience triple conjunctions. The GJ 876 system, by contrast, comes close to a triple conjunction between the outer three planets once per every orbit of the outer planet, "e."