Author(s): Santiago Guizzardi; Juan Bianchi; Julian Evaristo Cortese; Marcelo Uriburu Quirno; Martin Sabarots Gerbec

Linked Author(s):

Keywords: Parana Delta; Hydrologic forecast; Hydrodynamics

Abstract: The hydrodynamics of the Paraná Delta is highly conditioned by the interaction of the Paraná River discharge and the tidal regime of the Río de la Plata estuary. In 2020 and 2021, the Paraná River observed one of its most significant low-flow periods in its historical record and the most extreme after the river became regulated due to the construction of numerous dams in the upper Paraná basin (starting in 1975). The observed low-flows are having a significant impact on water intakes (for consumption and cooling of power plants) and the navigability of the waterway. A forecasting system has been implemented to predict water levels in different cross sections at two different lead times, i.e., 4 and 15 days, corresponding to different meteorological tide inputs, derived from wind forecasts. The system is based on a one-dimensional hydrodynamic model of the Paraná Delta, a two-dimensional model of the Río de la Plata estuary, wind input from two numerical weather prediction (NWP) models and autoregressive model schemes to further correct forecasts with observed data. The results are issued and updated every 6 hours, in interoperable format. The 4-day forecast model performance was assessed between 12/08/2021 and 25/1/2022 by means of the following metrics: mean absolute error (MAE), standard error (SE), and Pearson correlation coefficient (PCC), computed on modeled and observed series. For all lead times, an unbiased MAE was obtained. SEs of 0.21m for the 1-day lead time, gradually increasing to 0.29m for the 4-day lead time were obtained. Likewise, PCCs of 0.91 at 1-day lead time, decreasing to 0.82 at 4-day lead time were attained. In the case of the 15-day forecast model, the performance proved lower. The same performance metrics were calculated showing that the MAE was also unbiased. SEs go from 0.35m to 0.88m for 1 to 15 days of lead time, respectively, and the PCCs, from 0.73 to 0.22. This tool has had a significant impact among users in the region in terms of its outreach, as shown by the large number of website visits to the published forecasts. Even though an average of 130 daily visits are registered, there is uncertainty in the actual number of stakeholders that make use of this forecast system. It is estimated that the amplification factor is, at least, 3 to 4 times the web-access counts. This information dissemination takes place by less formal, yet effective, communication ways, as WhatsApp groups administrated by other government agencies or decision makers. In addition to the forecast system implementation, the present paper also shows high correlation between the demand for reliable information and the intensity of the extreme event, in this case, low water levels due to severe drought conditions in the upper basin.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221831

Year: 2022