Inicialização do modelo WRF pelo filtro digital

Abstract

WRF model initialization by digital filter The present work deals with the initialization of the WRF model by the digital filter. That technique is used to filter high-frequency gravity waves. High-frequency gravity waves arise due to the imbalance between the pressure and wind fields in the initial condition of the model and tend to spread through the domain. The causes of the imbalance between the pressure and wind fields are varied, such as: error in the observed data, imperfection of the numerical method that discretizes the differential equations by finite differences, grid resolution, difficulty to represent the nonlinear terms of the equations of the model and the truncation error in spectral models. In the generation of the filtered time series, oscillations occur close to the cut-off frequency (Gibbs oscillations), which are smoothed with the use of so-called window functions. In the proposed methodology, it is described the operation of the digital filter and the most efficient window function, according to a literature (Dolph-Chebyshev). Different cut-off frequencies, serial sizes, window functions and initialization shapes are also investigated to generate the filtered series. In addition, the results are explored through the simulation of a case of explosive cyclogenesis that occurred on 03/01/2014. The results demonstrate that the model simulated the cyclogenesis, when compared to CPTEC analyzes (used as reference). The digital filter showed an efficient methodology for the initialization of the WRF, with nested grids. The best results were obtained with1-hour cut-off frequency, 1-hour series, Dolph-Chebyshev window function and TDFI integration estrategy.