Assimilation of precipitation radars on TRMM and GPM Core satellites

Since the GPM Core observatory satellite successfully launched in February 2014, we have two valuable spaceborne precipitation radars of DPR and TRMM/PR. Assimilating those radars will be beneficial in numerical weather prediction (NWP) because they provide information on vertical precipitation profile over the both land and sea that are not available with other instruments. They, however, have been underused due to the high nonlinearity of precipitation-related variables and difficult precipitation reproductivity of NWP models.

An ensemble-based variational assimilation (EnVA) scheme has been developed to address the nonlinearity issue and effectively utilize precipitation-affected satellite data. The EnVA scheme uses a cloud-resolving model of the non-hydrostatic model of JMA (JMA-NHM) to explicitly treat with precipitation variables. It was developed to assimilate microwave imagers and has been expanded for precipitation radars. We investigated how well JMA-NHM reproduced TRMM/PR in the radar reflectivity space. It was found that JMA-NHM overestimated the reflectivity from ice particles and was not able to simulate bright bands. These results will help us to develop quality control procedures and specify observation error statistics that are used in the radar assimilation. Preliminary results of the model comparison and assimilation of PR and DPR will be presented.