Initialization with diabatic heating from satellite-derived rainfall

In this paper, a new technique is proposed to improve initialization of a tropical cyclone (TC) prediction model using diabatic heating profiles estimated from a combination of both infrared satellite cloud imagery and satellite-derived rainfall. The method is termed Rainfall-defined Diabatic Heating, RDH. To examine the RDH performance, initialization and forecast experiments are made with the Australia Bureau of Meteorology Research Centre (BMRC) Tropical Cyclone - Limited Area Prediction System (TC-LAPS) for the case of TC Chris, which made landfall on the west coast of Australia during 3-6 Feb 2002. RDH is performed in three steps: 1) based on previous observational and numerical studies, reference diabatic heating profiles are firstly classified into three kinds: convective, stratiform or composite types; 2) NRL (Naval Research Laboratory) 3-hourly gridded satellite rainfall estimates are categorized as one of the three types according to the rain rate; 3) within a nudging phase of 24 h, the model-generated heating at each grid point during the integration is replaced by the reference heating profiles on the basis of the satellite-observed cloud top temperature and rainfall type. The results of sensitivity experiments show that RDH has a positive impact on the model initialization of TC Chris. The heating profiles generated by the model within the observed rainfall area show agreement with that of reference heating. That is, maximum heating is located in the lower troposphere for convective rainfall, and in the upper troposphere for stratiform rainfall. In response to the replaced heating and its impact on the TC structure, the model initial condition and forecasts of the track and intensity are improved. (C) 2007 Elsevier B.V. All rights reserved.