Análise das séries vazões medidas em estações fluviométricas do estado do Ceará

Abstract

Ceará is located in Brazil’s semi-arid region, characterized by high rainfall variability, strong seasonality of runoff, and recurrent droughts, which makes understanding the hydrological regime essential for water resources planning and management; in this context, the longstanding reliance on reservoirs to regulate water supply is a central element. This study analyzed annual mean streamflow time series measured at streamgauge stations in the State of Ceará (Brazil), aiming to identify signals of nonstationarity (trends and abrupt changes) and to discuss their temporal compatibility with the presence of upstream reservoirs. Initially, hydrological data were obtained from an official database (ANA HidroWeb) and processed and organized in a computational environment, including the construction of a missing-data map to assess record completeness. Next, gauge stations with sufficiently long records and a low proportion of gaps were selected to increase the reliability of statistical inferences. To detect monotonic trends and quantify their magnitude, non-parametric methods were applied (Mann-Kendall test and Sen’s slope estimator). Abrupt changes in series behavior were investigated using the Pettitt changepoint test. Statistical routines were implemented in R (RStudio). In parallel, spatial analysis was carried out in a GIS environment (QGIS) to relate each station to its contributing watershed and to identify reservoirs located upstream, based on a public inventory monitored by COGERH. Finally, the change-point year estimated by Pettitt was compared with the commissioning/operation year of upstream reservoirs using a ±5-year tolerance window as a criterion for temporal coincidence, without implying causal inference. The results support the understanding of changes in Ceará’s hydrological regime and contribute to water-resources planning and management in a semi-arid region.