Análise espaço-temporal das precipitações máximas diárias anuais no estado da Paraíba

Abstract

The objective of this study was to analyze the spatiotemporal behavior of the annual maximum daily precipitation series in the state of Paraíba using observed data for the period 1962 to 2017. In addition to analyzing a series of 56 years of data (1962 – 2017), defined two 30-year time intervals of rainfall data for the periods 1962-1991 and 1988-2017 in order to study the behavior of maximum rainfall independently. The research consisted of five steps, namely: the selection of rainfall stations, the estimation of annual maximum daily rainfall for various return periods (eg, 2, 5, 10, 20, 50 and 100 years) using the probability distribution of Gumbel, the application of the Kolmogomoff - Smirnoff (KS) adherence test, and the spatialization of annual maximum daily rainfall using the ArcGis Pro software with the application of interpolation by Kriging and the use of the clustering technique to identify homogeneous rainfall zones through the k-means algorithm. As a result, the maximum annual daily rainfall estimated by the Gumbel probability function adhered well to the probability distribution at the 5% significance level. After spatializing the data, it can be inferred that the highest annual daily maximum precipitation values are found in the Sertão Paraibano mesoregion, which confronts common sense, but this result is justified by the state's average annual precipitation map, where it can It should be noted that as one advances towards the interior of the state, precipitation reaches higher intensities in the Alto Sertão Paraibano. Finally, the grouping technique was effective in the characterization of homogeneous pluviometric regions, as it presented similarity to the rainfall volume of Paraíba, and it can also be inferred that the altimetric level did not influence the formation of groups. Furthermore, rainfall variability in Paraíba was lower in Agreste Paraibano and higher in Sertão when considering the long series of data. In the subseries from 1962 to 1991, the pluviometric stations belonging to the Sertão Paraibano mesoregion also had the greatest dispersion in the data. In the subseries from 1988 to 2017, rainfall variability was lower in the Agreste Paraibano and higher in the Borborema mesoregion. Therefore, the opening of time intervals within the long series was of paramount importance to verify trends and make inferences.