This paper presents the results and analyses of Doppler radar measurements taken as part of a long-term campaign conducted in Singapore. Eight months of data taken in the zenith-pointing mode have been analyzed, in order to examine the validity of the assumptions underlying the retrieval algorithm proposed for the precipitation radar on board the TRMM satellite. Stratiform and convective events were analyzed separately. The former exhibits a clear melting layer in the 4--5 km height region, defined by an enhancement in the radar reflectivity, an increase in the mean fall velocity from 1--2 m/s in the snow region just above the melting layer to over 8 m/s in the rain region just below it. The corresponding width increases from less than 1 m/s to greater than 1 m/s. The variation in all three parameters can be accurately represented by the Noncoalescence-Nonbreakup melting layer model used for the satellite radar retrieval algorithms. Also examined was the monthly variation of the reflectivity-height profiles. In the rain region below the melting layer, a consistent decrease in the radar reflectivity with decreasing height is seen, indicating that the drop size distribution (DSD) cannot be assumed constant over all heights. However, if only the parameter N0 of the DSD is assumed constant, then good agreement is obtained between the mean velocity estimated from the reflectivity data and the measured velocity throughout the height profile. The agreement is also obtained in the spectrum width profile in the rain region. Convective precipitation on the other hand, does not exhibit a melting layer, and its averaged dBZ and velocity profiles cannot be characterized even under the constant N0 assumption. Its averaged reflectivity-height profiles show more variability than stratiform precipitation. For such cases, factors such as up and down drafts and nonstandard atmosphere need to be taken into account. The implication for the so-called N0 adjustment method is that while it can be applied to stratiform precipitation, the convective events need additional information, such as dual-frequency radar measurements, in order to improve the accuracy of the rainfall estimates.