The purpose of the present study is to retrieve latent heat profiles in hurricane from spaceborne passive microwave measurements. Then, the connection between hurricane's evolution and latent heat profiles will be presented.

A bayesian algorithm (BRAIN) is used to retrieve hydrometeors profiles from TMI brightness temperature. To support the inversion, we used an original database built with co-located data from TRMM-Precipitation radar (PR) and TRMM-Microwave radiometer (TMI).

Then, the latent heat release is expressed as the vertical fluxes of the different water species which include not only the water content but also some other information not directly retrieved by the BRAIN algorithm.

Among the variables we need is the vertical velocity. Thanks to the outputs of the meso-NH simulation model, we determined a simple relationship between the vertical velocity and hydrometeor profiles. Results of this regression and the correlation coefficients for each atmospheric layer will be presented. Since our retrieval method provides us the vertical hydrometeors profiles, we can now retrieve the vertical velocity. The last terms requiered are the hydrometeors terminal fallspeed velocities for which we use parameterised relations from the litterature. The latent heating calculation can now be performed using an approach described in Roux and Sun (1990).

The studied case is the tropical hurricane Bret (21-23 August 1999). Results from successive orbits will be presented. The diagnosed evolution will be compared to independant sources of data.