Example of a path of the HARPS fibers across the reconstructed Moon image, during the 300 s exposure in the penumbra. The darkening from the penumbra and umbra are not superimposed on the image here, since we are only interested in recovering the moon reflectivity. The A and B fibers remained at 232 and 125 arcsec from the umbra edge, while the fibers path across the Moon was ≈164 arcsec long (length of the yellow lines). The mean reflectivity is measured along the path, taking into account a 10 arcsec pointing error in both RA and Dec directions. The terrain reflectivity fluctuates by σ = 21% along the path for Fiber A and σ = 15% for Fiber B in this highly contrasted lunar region. The error on the mean reflectivity due to pointing uncertainty is estimated to be about σ = 4%. The figure shows that while Fiber A went through the dark region of Mare Nectaris, Fiber B remained on a region measured 39% brighter between Catharina and Cyrillus craters, pointing out the relevance of correcting the flux for lunar local reflectivity.