Towards generalized hydrological forecasting using transformer models for 120 h streamflow prediction
- 1University of Iowa
- 2ByWater Institute, Tulane University
- 3Tulane University
Abstract
This study explores the efficacy of a Transformer model for 120-hour streamflow prediction across 125 diverse locations in Iowa, US. Utilizing data from the preceding 72 hours, including precipitation, evapotranspiration, and discharge values, we developed a generalized model to predict future streamflow. Our approach contrasts with traditional methods that typically rely on location-specific models. We benchmarked the Transformer model's performance against three deep learning models (LSTM, GRU, and Seq2Seq) and the Persistence approach, employing Nash-Sutcliffe Efficiency (NSE), Kling-Gupta Efficiency (KGE), Pearson's r, and Normalized Root Mean Square Error (NRMSE) as metrics. The study reveals the Transformer model's superior performance, maintaining higher median NSE and KGE scores and exhibiting the lowest NRMSE values. This indicates its capability to accurately simulate and predict streamflow, adapting effectively to varying hydrological conditions and geographical variances. Our findings underscore the Transformer model's potential as an advanced tool in hydrological modeling, offering significant improvements over traditional and contemporary approaches.
Related
- XAI-driven analysis of influencing factors in streamflow-based flood forecasting via fine tuned MTV19Anet
- Advanced machine learning techniques for daily streamflow forecasting: a case study of the Brahmaputra River
- KNOU-Flood: Advancing global streamflow forecasting with lead-aware probabilistic machine learning
- Improved seasonal hydrological forecasting for Great Britain
- Machine Learning-Based Hydrological Drought Prediction Integrating Teleconnections and Hydrological Memory in a Semi-Arid Basin, Algeria
BibTeX
@article{demiray2026towards,
title = {{Towards generalized hydrological forecasting using transformer models for 120 h streamflow prediction}},
author = {Bekir Zahit Demiray and İbrahim Demir},
journal = {Machine Learning Earth},
year = {2026},
month = jun,
doi = {10.1088/3049-4753/ae81b8},
url = {https://doi.org/10.1088/3049-4753/ae81b8},
note = {CC BY},
}
RIS
TY - JOUR
TI - Towards generalized hydrological forecasting using transformer models for 120 h streamflow prediction
AU - Bekir Zahit Demiray
AU - İbrahim Demir
PY - 2026
DA - 2026/06/24
JO - Machine Learning Earth
DO - 10.1088/3049-4753/ae81b8
UR - https://doi.org/10.1088/3049-4753/ae81b8
AB - This study explores the efficacy of a Transformer model for 120-hour streamflow prediction across 125 diverse locations in Iowa, US. Utilizing data from the preceding 72 hours, including precipitation, evapotranspiration, and discharge values, we developed a generalized model to predict future streamflow. Our approach contrasts with traditional methods that typically rely on location-specific models. We benchmarked the Transformer model's performance against three deep learning models (LSTM, GRU, and Seq2Seq) and the Persistence approach, employing Nash-Sutcliffe Efficiency (NSE), Kling-Gupta Efficiency (KGE), Pearson's r, and Normalized Root Mean Square Error (NRMSE) as metrics. The study reveals the Transformer model's superior performance, maintaining higher median NSE and KGE scores and exhibiting the lowest NRMSE values. This indicates its capability to accurately simulate and predict streamflow, adapting effectively to varying hydrological conditions and geographical variances. Our findings underscore the Transformer model's potential as an advanced tool in hydrological modeling, offering significant improvements over traditional and contemporary approaches.
SN - 3049-4753
ER -