The parameterization of the microphysical characteristics for low-level stratiform water clouds can be developed in terms, among others, of the effective radius of droplets and the liquid water content (LWC). These parameters can be directly measured using aircraft mounted in-situ probes observations. The instruments used to perform these measurements, however, have an extremely small sample volume. Consequently, a large number of expensive and labour intensive flights are necessary to acquire statistically reliable profiles. The remote sensing methods are less direct but give much better coverage and are much less expensive. In this paper a retrieval technique based on the relationship between the effective radius of cloud drops and the radar reflectivity-to-lidar extinction ratio is presented. The existence and stability of such relationship using in-situ aircraft data for a few field campaigns that took place in different geographical regions, inside different cloud types, and under different meteorological conditions are demonstrated. The data for the CLARE'98 (October 1998, Chilbolton, UK, stratiform clouds), the DYCOMS-II (July 2001, Pacific Ocean near California coastal zone, stratiform clouds), the CAMEX-3 (August - September 1998, Florida, clouds in tropical storm) campaigns were analyzed. For all these field campaigns the unified procedure for calculation of the drop size distributions (DSD) from the measured with different probes data was used. The resulting merged distributions were used for the calculation of the "effective radius - to - radar reflectivity-to-lidar extinction ratio" scattering diagrams and two-dimensional histograms. The analysis of the data for different campaigns shows a good agreement between their behavior on such plane and the possibility to use a joint distribution over all field campaigns for estimation of the mean "effective radius - to - radar reflectivity-to-lidar extinction ratio" relationship. The difficulties of the parameter estimation for piecewise-linear fitting of the mean "effective radius - to - radar reflectivity-to-lidar extinction ratio" relationship were shown and the possibility to use unified for all analyzed campaigns and cloud types 4th order polynomial fitting was demonstrated. The theoretical analysis of the statistical models for the observed DSDs shown that the information about DSD position on the "effective radius - to - radar reflectivity-to-lidar extinction ratio" plane can be used for cloud type classification into three classes: clouds without drizzle, clouds with drizzle, and drizzle clouds. The application of the developed classification technique to the representation of observed data on the LWC-Z plane allowed us to use the different LWC-Z relationships for different cloud types. The previously published relationships for clouds without drizzle; for cloud with drizzle; and the new relationship for drizzle clouds show reasonable good agreement with observed data. The applicability of the remote sensing technique based on simultaneously measured radar and lidar data for the cloud's type classification and selection of the LWC-Z relationship is demonstrated and validated using real observations with radar, lidar, and radiometer.