On the comparison between presumed and full pdf methods for turbulent precipitation