Article ID: | iaor19991867 |
Country: | Brazil |
Volume: | 16 |
Issue: | 2 |
Start Page Number: | 77 |
End Page Number: | 87 |
Publication Date: | Dec 1996 |
Journal: | Pesquisa Operacional |
Authors: | Alves J.M.B., Ferreira N.S., Repelli C.A. |
Keywords: | time series & forecasting methods |
This paper presents a trend analysis of a number of years with rainfall periods classified as Very Dry (VD), Dry (D), Normal (N), Rainy (R) and Very Rainy (VR) to some rainfall data sets from the north sector in the north-east of Brazil for the period of 1990 to 2040. This trend was based on the rainfall climatic characteristics observed in this century. The technique used to extrapolate it was Fourier Series (FS). Performed tests have evidenced that the extrapolation by FS could be utilized to quantify the number of rainfall periods, classified in determinate categories, to a period defined before without considering their interannual variabilities. The results show that the trend of the sum of the years classified as VD and D categories, years that cause more negative effects in the socio-economic aspect of the region, did not exceed in any sub-region studied the sum of N, R and VR categories. These, in theory, are more favorable to these aspects, in particular, to the accumulation of water in reservoirs.