Assessing Return Periods of New Extreme Events: Preliminary Analysis of Peak Flows in the May 2024 Hydrological Event in Rio Grande do Sul, Brazil

Código

I-EBHE0169

Autores

Maurício Cezar Rebello Cordeiro, ALAN VAZ LOPES, Paulo Roberto Roballo Ungaretti, MARCOS AIRTON DE SOUSA FREITAS, BRUNA CRAVEIRO DE SA E MENDONCA

Tema

WG 1.08: Deep Explanation & Evaluation for Practices in Hydrological Changes (DEEPHY)

Resumo

The extreme rainfall events that occurred in April and May 2024 over Rio Grande do Sul State raised river levels above historically observed maxima, causing significant impacts across several cities in the Taquari, Jacuí, Caí, and Uruguai river basins. The extreme flows resulted in structural damage to bridges and dams, suggesting that design flows might have been exceeded in some cases. This raised questions about the actual return periods of the 2024 floods and which design flows should be used in the reconstruction of the damaged infrastructure. However, estimating the return period of those events is challenging because the recent floods might have been influenced by climate change and their return might not be accurately described by probabilistic models based on historical observations. To investigate possibilities and limitations in estimating return periods of extreme flood events using probabilistic models and historical records, this study presents a preliminary analysis of the peak flows observed during this extreme hydrological event. Considering data availability, a total of 10 hydrological stations were included in the study. Historical streamflow time series were obtained from the National Water and Sanitation Agency (ANA) database, with series lengths ranging from 31 to 83 years, while peak flows during the event were obtained from existing stage-curves and observed water level data from different sources, according to the reality in the field. The recurring time analysis employed empirical probability distributions and probabilistic adjustments using Log-Normal and Gumbel distributions through an automated process. To test the adjustment of the probabilistic distributions, the Kolmogorov-Smirnov (KS) test was applied to each station and its results compared to its critical value (Dc). Despite the severity of the floods, only 5 out of 10 stations had an all-time-high discharge flow during the 2024 event. The higher return periods were obtained in Jacuí River at Barra do Caí and Rio Pardo stations, with values on the order of 800 years and 2,000 years, respectively. In the other stations, the return periods varied between 35 to 113 years. However, in 4 stations, the KS test or even visual inspection indicates that conventional probability distributions were not suitable for estimating the return period of the maximum flows observed in the 2024 flood event. Considering also the recent increasing trend in observed maximum flows, our analysis suggests that, in some locations, climate change might have contributed to increasing the magnitude of the observed peak flows. Therefore, new methods and approaches are urgently needed to adequately estimate return periods and risks associated with extreme floods in the region. Finally, the obtained results may shed light on possible future approaches and provide insights into the return periods of extreme hydrological events in the region, helping in the design of new infrastructure projects.