To understand the frequency, and thus the formation and evolution, of planetary systems like our own solar system, it is critical to detect Jupiter-like planets in Jupiter-like orbits. For long-term radial-velocity monitoring, it is useful to estimate the observational effort required to reliably detect such objects, particularly in light of severe competition for limited telescope time. We perform detailed simulations of observational campaigns, maximizing the realism of the sampling of a set of simulated observations. We then compute the detection limits for each campaign to quantify the effect of increasing the number of observational epochs and varying their time coverage. We show that once there is sufficient time baseline to detect a given orbital period, it becomes less effective to add further time coverage -- rather, the detectability of a planet scales roughly as the square root of the number of observations, independently of the number of orbital cycles included in the data string. We also show that no noise floor is reached, with a continuing improvement in detectability at the maximum number of observations N=500 tested here.
We report the detection of two new planets from the Anglo-Australian Planet Search. These planets orbit two stars each previously known to host one planet. The new planet orbiting HD 142 has a period of 6005pm427 days, and a minimum mass of 5.3M_Jup. HD142c is thus a new Jupiter analog: a gas-giant planet with a long period and low eccentricity (e = 0.21 pm 0.07). The second planet in the HD 159868 system has a period of 352.3pm1.3 days, and m sin i=0.73pm0.05 M_Jup. In both of these systems, including the additional planets in the fitting process significantly reduced the eccentricity of the original planet. These systems are thus examples of how multiple-planet systems can masquerade as moderately eccentric single-planet systems.
