Towards Generalized Hydrological Forecasting using Transformer Models for 120-Hour Streamflow Prediction

TL;DR

This paper demonstrates that Transformer models can effectively predict 120-hour streamflow across diverse locations, outperforming traditional deep learning models and benchmarks, thus offering a generalized approach to hydrological forecasting.

Contribution

The study introduces a generalized Transformer-based model for streamflow prediction that works across multiple locations, contrasting with traditional location-specific hydrological models.

Findings

Transformer outperforms LSTM, GRU, Seq2Seq, and Persistence models.

Transformer achieves higher NSE and KGE scores.

Transformer exhibits lower NRMSE, indicating better accuracy.

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.