A widely used method to gain a better estimation of ground precipitation is to fit a vertical profile of reflectivity (VPR) to volume radar measurements. In this work good resolved VPRs were derived from range height indicator measurements. After a necessary scaling procedure they were statistical analyzed using empirical orthogonal functions (EOFs). A neuronal network was trained to guess the according EOF coefficients from a small number of fulcrums corresponding to volume scan measurements. Guessing these coefficients makes it possible to reconstruct the whole VPR from volume scan measurements and thereby a better estimate of ground precipitation.