Reconhecendo eventos hidrológicos: indissociando magnitude extrema, duração e período de ocorrência na análise de frequência (raridade)

Abstract

With the need to analyze the aspects of events with methods that present fewer uncertainties. This work aims to propose a new approach in order to reduce the uncertainties in the analysis of hydrological events considering the different variables of an event as interdependent, thus obtaining more information about the events. The study was applied to the São Francisco River basin, the river of national integration that suffers from different anthropic actions, from the Xingó station with observed flow data provided by the Companhia Hidrelétrica do São Francisco (CHESF)/SAR of ANA, and natural data reconstituted by the Operador Nacional do Sistema Elétrico (ONS), the data are available for the period 1976-2018 for observed data and 1931-2018 for natural data. Event analyses usually consider one variable, maximum or minimum magnitude per year. Currently, the use of copula functions is increasing in multivariate analysis of hydrological events, as it is known that it is necessary to consider more information about the events. In Hydrological Alteration Indicators (HAI), a method to obtain information about alteration of a regime by anthropic actions, the frequencies of the aspects (aggregation of variables) of an event are analyzed considering the independent variables. However, in this work, copula functions were applied to analyze the aspects (aggregating variable) of an event considering its variables as being dependent. The variables flow, duration and period of occurrence are the ones chosen for this work, and the marginal distributions for each variable, respectively, Generalized Pareto (GP), Pearson 3, and Von Mises (floods) and Normal (droughts). The Gaussian copula showed the best results in the goodness of fit tests for modeling the flood and drought events, while the Archimedean copulas had satisfactory results for the flood events and unsatisfactory results for the drought events. A new approach for hydrological regime change analysis was introduced, thus showing changes in the flow regime of the São Francisco River.