On the impact of Doppler radar derived wind fields in a mesoscale non-hydrostatic model

Fine-scale three-dimensional wind fields retrieved from airborne Doppler ra dar observations within an oceanic tropical mesoscale convective system (MC S), during the Tropical Ocean/Global Atmosphere Coupled Ocean-Atmosphere Re sponse Experiment, are used to enhance the initial conditions of a non-hydr ostatic mesoscale model. They are incorporated into the French CANARI-ALADI N analysis system, based on the optimal interpolation technique. The fine-m esh (4 km) horizontal wind components are first averaged over subdomains of 20 km x 20 lull in order to provide a series of mesoscale wind profiles. B ecause of the absence of dense sounding data, the synthesized vertical-velo city profiles are transformed into humidity profiles by considering that up draughts contain nearly saturated or saturated air, while downdraughts are associated with unsaturated air. Comparison of the initial state using conventional data with the Doppler-en hanced initial state clearly identifies the benefits of the addition of the observed mesoscale circulation features. Ln terms of the precipitation for ecast, the control run that uses the conventional initial state dramaticall y fails to predict the existence of the MCS precipitation core. On the othe r hand, the reference sun with the: data-enhanced initial state succeeds in forecasting it up to 12 hours in a way that is well consistent with the sa tellite imagery. In terms of the wind forecast, the mesovortex signature th at could be identified from the radar observations becomes a persistent fea ture with the radar-enhanced initial conditions. Sensitivity tests reveal t he specific roles of wind and humidity data. As in previous studies, the in itial stare of humidity is fundamental for the forecast of the MCS precipit ation by sustaining the convective activity, and also by forming and mainta ining a vortex-like circulation. These results suggest that moist convectiv e processes play a major role. The inclusion of wind data is necessary for improving the system propagation by advective processes.