Monthly precipitation in southeast Brazil - for analysis of random and trend changes

Abstract

In this work, statistical analysis tools of the computational program were used: the computational statistical package R. Data from the monthly precipitation series were analyzed in five airports. It was shown that monthly rainfall showed variability associated with three distinct subgroups, with different intensities of precipitation and represent an overlap in cascades of different spatial and temporal scales, acting on the pattern of precipitation behavior. The use of the chi-square test using as reference the main maximal centroids, obtained by the group analysis technique, allows characterizing the variability of the maximal and possible analysis of the frequency change in the behavior of the time series. In the Spearman Coefficient test, he presented two locations with a trend, Guarulhos and Santa Cruz. However, considering the results and analyzing the metadata information from the observation stations on the surface, we can consider that there is evidence that the variability is within randomness and without showing changes in trends.