Comparing MLP and LSTM neural networks for short-term water level forecasting: insights from the 2024 major flood in Southern Brazil
Open Access
- 1 January 2026
- journal article
- Published by FapUNIFESP (SciELO) in RBRH
- Vol. 31
Abstract
This study develops and compares two data-driven models for short-term water level forecasting based on a MLP and a LSTM network. Both models were trained for a six-hour forecasting horizon using hourly water level and rainfall data from seven stations in the Guaíba Lake watershed, southern Brazil, from November 2023 to May 2024, a period that includes the region’s most severe flood on record. A sliding-window learning scheme with incremental weight updates was employed to improve temporal consistency and computational efficiency. Four experiments assessed the effects of different input configurations, recursive forecasts, and model architectures. Results indicate that the MLP performed comparably to, and slightly better than, the LSTM, with MAE and NSE values supporting its suitability for near-real-time forecasting. Overall, the findings show that relatively simple neural architectures, when trained with adaptive schemes, can deliver robust and efficient predictions for operational early-warning systems.Keywords
Funding Information
- CNPq (446053/2023-6)
- CNPq (305205/2025-0)
- CNPq (2024/02748-7)
- FAPESP (446053/2023-6)
- FAPESP (305205/2025-0)
- FAPESP (2024/02748-7)
This publication has 21 references indexed in Scilit:
- Floods in 2025Nature Reviews Earth & Environment, 2026
- Machine learning-based hydrological models for flash floods: a systematic literature reviewSmart Construction and Sustainable Cities, 2025
- A machine learning-based prediction-to-map framework for rapid and accurate spatial flood predictionnpj Natural Hazards, 2025
- Unprecedented flooding in Porto Alegre Metropolitan Region (Southern Brazil) in May 2024: Causes, risks, and impactsJournal of South American Earth Sciences, 2025
- ML4FF: A machine-learning framework for flash flood forecasting applied to a Brazilian watershedJournal of Hydrology, 2025
- Nonlinear hydrological time series modeling to forecast river level dynamics in the Rio Negro Uruguay basinChaos: An Interdisciplinary Journal of Nonlinear Science, 2024
- Uncovering Flooding Mechanisms Across the Contiguous United States Through Interpretive Deep Learning on Representative CatchmentsWater Resources Research, 2022
- An operational dynamical neuro-forecasting model for hydrological disastersModeling Earth Systems and Environment, 2016
- Out-of-sample tests of forecasting accuracy: an analysis and reviewInternational Journal of Forecasting, 2000
- Artificial Neural Networks in Hydrology. II: Hydrologic ApplicationsJournal of Hydrologic Engineering, 2000