Avaliação dos padrões de variabilidade e mudança climática no setor hidrelétrico brasileiro

Abstract

This sduty aims to evaluate possible trends, coupled with climate change, in naturalized flows time series iat the stations of the Operador Nacional do Sistema (ONS). The analytical methods used were the so-called classical methods (mean and median mobile, linear regression, LOWESS and Mann-Kendall-Sen), the study of peak floods by adjusting the time series to Gumbel distribution, analysis of wavelet transforms and the correlation between flows and climate indices. The results showed tha there is a positive trend or a possible low-frequency variability in south, southeast and midwest of Brazil, while in the northwest, there is a negative trend or low frequency variability in flows. It was also found a coincidence between the step change of flow regime in some stations with high values of PDO index, wiich may indicate tha there is a low frequwncy variability in flows. The analysis of wavelet transformation indicate that there is a low frequency of variability with a period between 40 and 80 years and an average frequency of variability with a period between 10 and 20 years in most flow stations analyzed. The results for the CMIP3 climate models indicate that, in the IPCC A1B scenario for the 21th century, there is an increase in the variability of flows over the twentieth century in the most of models of the 24 selected stations of SIN for this study. The results of the correlation between the average monthly flow and the climate indices, showed that only Niño 1+2 index have a moderate linear correlation.