In the field of atmospheric research, ground-based radar systems are often employed to study ice/mixed-phase cloud properties based on retrieval techniques. These techniques convert the radar signal backscattered by each bulk of ice crystals being probed within the same radar resolution volume to cloud’s microphysical characteristics. However, the size of a radar resolution volume is often too large compared to the microphysical and thermodynamical variability of the atmosphere. The microphysical information contained in the radar signal is then complex and difficult to retrieve. Therefore, the ground-based observations of the cloud’s particles microphysical characteristics are a real challenge in atmospheric science. In this thesis, an advanced atmospheric profiling radar (TARA - IRCTR) simultaneously uses Doppler and polarimetric capabilities in order to characterise the microphysical properties of the ice crystals present in ice/mixed-phase clouds. On the one hand, the polarimetric response of an atmospheric target is related to its axis ratio, assuming a spheroidal shape. On the other hand, the Doppler effect induced by the particle motion is measured by the radar and is referred as particle radial velocity. Therefore, compared to conventional weather radars, Doppler-polarimetric measurements better describe the radar resolution volume of ice particles which are categorised according to particle shape and velocity (see Chapter 3). The main objective of this PhD thesis is to develop a new type of retrieval technique, where the microphysical characteristics of ice / mixed-phase clouds are obtained from the radar spectral polarimetric measurements. High data quality is required for the reliability of spectral polarimetric parameters. Thus, a specific signal processing is performed for each radar cell, so that noise and clutter free radar parameters are obtained. Moreover, the atmospheric instability effect on radar signal data has to be removed by correctly averaging the signal over time. A strong point of this thesis comes from the derivation of averaging on a radar cell basis, based on a statistical study of the spectral polarimetric parameters determining the time slots of the atmospheric instability (see Chapter 4). In Chapter 5, a microphysical model for the study of ice crystals conversion into raindrops was improved in order to relate the ice crystals properties of ice/mixed-phase cloud regions to spectral polarimetric data. The model is based on different relations which take into account the particle habit (assuming spheroidal shapes) and orientation, ice crystals’ maximum size and their size distribution (assuming that particles fit in a modified gamma distribution). The model for ice/mixed-phase cloud study was enhanced, following these three points: 1) the implementation of the column-like pristine ice, 2) the implementation of the vertical orientation of the ice particles; and 3) a full sensitivity analysis of the input as well as model related parameters. The major breakthrough of this thesis is described in Chapter 6 with the development of a new microphysical retrieval technique. For the first time, a detailed microphysical analysis of the ice crystals present in ice/mixed-phase clouds is obtained, at each radar cell, from the sole use of TARA radar (Transportable Atmospheric Radar) Doppler polarimetric measurements. The spectral polarimetric parameters are first used to determine the type of ice particles present in the radar resolution volume, relating to their main orientation, main size and habit. The model described in Chapter 5 is then applied to the retrieval technique, as a forward model, in order to determine the mean ambient radial wind velocity and the three free parameters of the modified gamma distribution referring to each particle type. The retrieval technique was tested for a specific meteorological condition during the COPS (Convective and Orographically-induced Precipitation Study) measurement campaign, in summer 2007. As seen in Chapter 7, the microphysical results and the interpretation of the cloud processes obtained so far, within a convective nimbostratus cloud, showed good spatial and time regularity. The radar measurement results were compared and validated with other collocated sensors. The comparison of the microphysical cloud properties of a similar cloud condition was achieved with collocated aircraft measurements, flying over the ground-based site. The mean Ice Water Content (IWC), the mean ice Total Number Concentration (Nt) and the shape of the PSD, retrieved and measured by the ATR42 aircraft, were found in good agreement.