Precise orbits of the Lunar Reconnaissance Orbiter from radiometric tracking data

Since 2009, the Lunar Reconnaissance Orbiter (LRO) acquires images and altimetric profiles of the lunar surface. Assembling these data to maps and terrain models requires the precise knowledge of the spacecraft trajectory. In this contribution, we present 5 years of LRO orbits from radiometric data processed with a software tailored to this mission. The presented orbits are the first independent validation of the LRO science orbits from NASA and are available for public use. A key feature of our processing is the elaborate treatment of model and observation errors by empirical parameters and an adaptive data weighting by variance component estimation. The quality of the resulting orbits is assessed by analyzing overlapping arcs. For our solution based on arcs of 2.5 days, such analysis yields a mean error of 2.81 m in total position and 0.11 m in radial direction. It is shown that this result greatly benefits from the adaptive data weighting, reducing the error by 2.54 and 0.13 m, respectively. Unfortunately, the precision achieved varies strongly, dependent on the view onto the orbital ellipse which changes with the lunar cycle. To mitigate this dependency, the arc length was extended in steps up to 10.5 days, leading in the best case to a further improvement of 0.80 m.