Generalized Dynamic Normal Mode Initialization
During the early phase of the numerical weather forecast, the initial analysis of the divergence plays an important role. However, using the conventional data analysis schemes, a successful initial analysis of the divergent circulation is difficult to achieve. It is proposed that over the convective regions, incorporating known convective heating information during the initialization can lead to an improved initial specification of the divergent circulation. Success of the proposed methodology depends on the efficiency of the initialization scheme. For the dynamic normal mode initialization (DNMI) method used in the present study to perform the initialization, it is shown that this method does not succeed in initializing higher vertical modes. Since the assimilation of the convective heating into the initialized divergence crucially depends on the ability to initialize the correct vertical modes, use of DNMI would not be very appropriate. An improved initialization procedure where the DNMI is performed in the vertical mode space was then formulated. A comparison of initialization experiments performed with the conventional DNMI and the new scheme demonstrated that with the new initialization method, the initialization of the higher vertical modes could be better achieved. The effects of including the convective heating during the initialization on the final initialized state were next explored. It is demonstrated that the dominant effect of the convective heating would be on the initialized divergence field. Further, it was illustrated that to some extent, the initialization with the specified convective heating can recover the corresponding divergent circulation in the initialized state. The assimilation of the convective heating information into the initialized divergence was next put into use in a more practical setting. Using the satellite derived rainfall rates, it was confirmed that the initialization resulted in an improved structure of the divergence field associated with a typhoon. The improvements in the initialized divergence also reflect in other initialized fields (e.g. moisture convergence, initial rainfall). In the subsequent forecast, externally specified convective heating used during the initialization was merged with the model generated heating to keep the initialized divergence still in balance. It is shown that then the improvements in the initialized divergence also have positive impact on the subsequent forecast.
- Pub Date:
- January 1990
- WEATHER PREDICTION;
- Physics: Atmospheric Science