Rainfall Distribution in the Brazilian Amazon: Application of the Variogram Function to Time Series

Abstract

The Amazon area is vast and diversified, with several climatic zones that influence directly in the associated ecosystems. This research analyzed the annual (hydrological year) and seasonal (drought and wet) climatological patterns of the Brazilian Amazon (North Region) between 1991–2022. The incomplete Gamma distribution model, kriging and cokriging interpolation tests, estimates of mean precipitation by the isohyet method, and of precipitated volume by the contour method were applied to numerical data. Data normalization techniques and image interpolation also contributed to identify precipitation patterns and trends in the northern region of Brazil. The method of hierarchical grouping by minimum variance was used, aided by the construction of frequency histograms and tests of significance. The results indicated a strong seasonality of rainfall. The highest precipitation rates occurred in the N-NW axis and the lowest in the south equatorial zone. The active atmospheric systems played an important role in the distribution of rainfall, both annual and seasonal. Four distinct climatic zones were established: southern, central, northwestern and northeastern. Based on the frequency of rainfall determined at levels of 10 to 90% probability, the results suggested a trend in the number of days with precipitation between 25 and 50% in most of the southern and central zones and an increase in precipitation from 50 to 75 % in the northern region. The ordinary kriging test and the spherical semivariogram model showed the best results for data interpolation, with better responses to the behavior of climatic variables and solutions for the spatial-temporal discontinuity of historical series. The results indicated a tendency towards the presence of isotropic climatic phenomena, especially in the wind distribution model. The results showed that they could support the understanding of the distribution of rainfall in the Amazon region.