Evaluation and diagnosis of physical processes in GCMs with improved hindcast (CAPT) approach

Description

We will present analyses of CAPT simulations with a refined hindcast procedure, and compare the results to ARM and other available observations. The motivation for the refined procedure is to minimize the errors in the initial conditions and allow for better evaluation of parameterizations' performance in the hindcast mode. For atmospheric initial conditions, we applied a nudging method in which only horizontal velocity fields are nudged towards the operational analysis/reanalysis. This way, a better representation of aerosols and cloud fields are obtained for the atmospheric initial conditions compared to observations. For land initial conditions, we obtained them from an offline land model simulation forced with observed atmospheric precipitation, winds, and surface fluxes. This way produces better soil moisture in the land initial conditions and better surface air temperature and fluxes in the hindcasts, especially over the mid-latitude continents during the summertime.