Abstract
The investigation of the dynamics of climatic variables provides important information about their spatio-temporal variability. Understanding these processes is fundamental for the development of climate models that serve as a basis for analysis of future scenarios and prediction of climate change. In this work, we analyze multifractal properties of daily air temperature series in Northeast Brazil recorded during the period 1990 to 2015, in 94 meteorological stations distributed over the region. We applied the Multifractal Detrended Fluctuation Analysis (MF-DFA) method that was developed to quantify long-range correlations in multifractal time series. The results show that the process that generates the variability of the air temperature follows the multifractal dynamics, with greater influence of the seasonal component. A persistent auto-correlation was observed, with stronger persistence in the proximity of the equator. The degree of multifractality indicated by the width of the multifractal spectrum varies according to the climatic classification of the region. The strongest multifractality was observed in the Mata and Agreste Zone. The results indicate that long memory in time series of the daily average air temperature of the Brazilian Northeast should be modeled as a multifractal process and may contribute to the development of more reliable regional climate models.
Keywords:
iimes series; Multifractal processes; MF-DFA
