Please use this identifier to cite or link to this item:
|Title:||Impact of digital filtering initialization on the performance of a semi-Lagrangian semi-implicit model over Indian region|
|Abstract:||Non-recursive digital filtering initialization (DFI) scheme has been applied adiabatically as well as diabatically to a semi-Lagrangian semi-implicit (SLSI) limited area model to study its impact on the model performance for short-range prediction over Indian region. For comparison purpose, the input is also initialized by dynamic normal mode initialization (DNMI). It is found that both the versions (adiabatic and diabatic) of DFI could produce fairly well balanced initial state and could effectively suppress the high-frequency oscillations from the forecast fields. The filter is found to be more effective in controlling the noise in the first few hours of integration. The changes induced to the analysis by both the versions o f DFI are reasonably small and are comparable with the changes made by DNMI. The 24-h forecasts obtained from the adiabatic as well as diabatic versions of DFI are comparable with those produced from DNMI and also with the forecasts obtained from uninitialized input. The main advantage of DFI is found to be its great simplicity in conception and application. Its implementation is very easy, requiring only the calculation of the filter coefficients and minor adjustments of the model code.|
|ISSN:||0975-105X (Online); 0367-8393 (Print)|
|Appears in Collections:||IJRSP Vol.29(6) [December 2000]|
Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.