Terminal fall velocity and the FASTEX cyclones

Knowledge of water‐drop and ice‐crystal terminal velocities is particularly important for an adequate representation of particle sedimentation in cloud‐resolving and large‐scale models. In this paper, a new method is proposed to retrieve terminal fall velocity from airborne‐Doppler‐radar observations. To extract the terminal fall velocity from the Doppler information, which mixes the three wind components and the terminal fall velocity, statistical considerations are introduced, stating that for a long sampling time span (a whole aircraft mission, for instance) and for moderate storms such as stratiform precipitation, the mean vertical air motions should vanish with respect to the mean terminal fall velocity. Statistical power‐law relationships between terminal fall speed and radar reflectivity are then developed, as well as profiles of mean terminal fall velocity. This underlying hypothesis of the method is validated with in situ measurements of the vertical wind component, averages of convective‐scale retrievals of the vertical wind component, and terminal fall speed–radar reflectivity ( V T– Z ) relationships derived from in situ measurements of the drop size distribution. These comparisons all indicate that the vertical air motion can indeed be neglected with respect to the terminal fall speed. A detailed analysis of the statistical relationships obtained in liquid and ice phases for six frontal cyclones at different stages of development shows that a ‘universal’ V T– Z rain relationship can be proposed for North Atlantic frontal cyclones at mature stage. This result and further arguments developed in the paper suggest that the drop size distributions are ‘equilibrium’ distributions. In the ice phase, such a ‘universal’ relationship is not found. It is nevertheless suggested in this study that such a general relationship might be derived if the frontal cyclones are split into categories depending on their stage of development. These V T– Z ‘universal’ relationships can be introduced in large‐scale model parametrization schemes in order to better describe the sedimentation of ice and water and dynamical–microphysical interactions occurring within North Atlantic frontal cyclones. Finally, a procedure is proposed to estimate the most representative type of ice particle from 2D‐P in situ sensor and airborne‐Doppler‐radar measurements. It consists of calculating the mean flight‐level Z and the V T– Z relationships in the ice phase for several theoretical ice habits, and to identify the ice habit that reasonably reproduces both the radar‐derived mean flight‐level Z and V T– Z ice relationship. Copyright © 2003 Royal Meteorological Society