THE IMPACT OF CLOUD PARAMETERS ON THE SIMULATED ERRORS IN RTTOV5

TOVS/ATOVS data from meteorologic satellite can provide the vertical distribution of temperature and humidity in the atmosphere. In order to use these data in NWP models directly, then to make the initial field of NWP model more accurate and improve forecast finally, a fast transmittance model (RT model) is required in a multi-dimensional data variational assimilation system. There will be big errors while simulating HIRS data by an RT model, because HIRS data in TOVS/ATOVS can not get the information of the vertical distribution of temperature and humidity below the clouds. Before assimilating TOVS/ATOVS data into NWP model, it is a necessary task to analyze the impact of the cloud parameters (the top of cloud and cloud fraction are included mainly) on simulated errors using RTTOV5, based on the actuality of operational TOVS data every day (about 70% observation points are partly cloudy).The temperature and humidity profiles from operational TOVS data, as well as some others data necessary to run RTTOV5, are used to compute the brightness temperature with respect to HIRS channels in NOAA14. Then the impact of the cloud fraction and the top of cloud on simulated errors in models are estimated. Finally an adjoint model and a Jacobine model of RTTOV5 are used to analyze the sensitivity of the simulated errors of RTTOV5 to initial cloud parameters (the cloud fraction and the pressure of of cloud top).