Intraseasonal oscillations are factors controlling the interannual variability of tropical rainfall. The knowledge of how numerical models reproduce its variability is important to better understand their roles and support operations of weather and climate. This paper investigates the sensitivity of a downscaling dynamic precipitation model in predicting the intraseasonal oscillations observed in the northern sector of Northeast Brazil (SNNEB - 45°W-37ºW and 2ºS-12°S) during the period from 1974 to 2000. The results showed that rainfall was simulated on SNNEB explained more than 70% of the variability of rainfall observed in the quarter February to April in this region. Statistical of averages, standard deviations and percent this quarter, in this region in years of climatic contrasts in the Oceans Pacific and Atlantic Tropical were also well captured by the precipitation simulated by downscaling. Spectral analysis with the use of wavelets showed that dynamic downscaling has the potential to reproduce observed spectral peaks of precipitation in SNNEB scales up to 8 days and the scales between 64 and 128 days for the period January 1 to June 30.
Intraseasonal rainfall; wavelets; spectral analysis
