Inversion methods of three cumulus parameterizations for diabatic initialization of a tropical cyclone model

[...]Aonashi ( 1993) developed a physical initialization scheme to enhance syn optic analyses of temperature, moisture, and divergence by incorporating precipitation observations from a radar-rain gauge network. Since the moist convective process is most important among the physical processes in a prediction model, as far as the amelioration of precipitation spinup is concerned, Donner (1988) proposed a scheme of "cumulus initialization" in which the temperature and moisture fields are adjusted such that a version of the Kuo cumulus parameterization will yield the precipitation rate close to "observed" when the input largescale vertical velocity supports such cumulus convection. In this paper we investigate inversion methods for three cumulus parameterizations in a regional model for prediction of tropical cyclones.