Assessing Reduced‐Dynamic Parametrizations for GRAIL Orbit Determination and the Recovery of Independent Lunar Gravity Field Solutions

Orbit determination of probes visiting Solar System bodies is currently the main source of our knowledge about their internal structure, inferred from the estimate of their gravity field and rotational state. Nongravitational forces acting on the spacecraft need to be accurately included in the dynamical modeling (either explicitly or in the form of empirical parameters) not to degrade the solution and its geophysical interpretation. In this study, we present our recovery of NASA GRAIL orbits and our lunar gravity field solutions up to degree and order 350. We propose a systematic approach to select an optimal parametrization with empirical accelerations and pseudo‐stochastic pulses, by checking their impact against orbit overlaps or, in the case of GRAIL, the very precise inter‐satellite link. We discuss how parametrization choices may differ depending on whether the goal is limited to orbit reconstruction or if it also includes the solution of gravity field coefficients. We validate our setup for planetary geodesy by iterating extended lunar gravity field solutions from pre‐GRAIL gravity fields, and we discuss the impact of empirical parametrization on the interpretation of gravity solutions and of their error bars. Plain Language Summary We analyze the orbits of the twin GRAIL probes orbiting the Moon in 2012 to determine the lunar gravity field. The frequency shift of radio signals exchanged between GRAIL and Earth antennas, as well as an ultra‐precise inter‐satellite radio link, allow for an accurate determination of the absolute and relative spacecraft trajectories. As these trajectories mainly depend on the internal mass distribution of the Moon, their accurate knowledge provides important information about lunar gravity and internal structure. However, nongravitational forces acting on the probes, for example, caused by radiation from the Sun or the Moon, have to be modeled as well. An explicit modeling of these forces requires an accurate knowledge of spacecraft geometry and optical properties, which are often poorly known. Additional accelerations can be estimated from the data to absorb the effect of nongravitational forces and improve both orbit and gravity field determination. We propose a systematic approach to configure such estimated accelerations when dealing with the recovery of planetary probes' orbits and gravity fields, while also discussing collateral effects of employing them. Based on these findings we also present our independent s