Abstract
Abstract. Long short-term memory (LSTM) models are recurrent neural networks from the field of deep learning (DL) which have shown promise for time series modelling, especially in conditions when data are abundant. Previous studies have demonstrated the applicability of LSTM-based models for rainfall–runoff modelling; however, LSTMs have not been tested on catchments in Great Britain (GB). Moreover, opportunities exist to use spatial and seasonal patterns in model performances to improve our understanding of hydrological processes and to examine the advantages and disadvantages of LSTM-based models for hydrological simulation. By training two LSTM architectures across a large sample of 669 catchments in GB, we demonstrate that the LSTM and the Entity Aware LSTM (EA LSTM) models simulate discharge with median Nash–Sutcliffe efficiency (NSE) scores of 0.88 and 0.86 respectively. We find that the LSTM-based models outperform a suite of benchmark conceptual models, suggesting an opportunity to use additional data to refine conceptual models. In summary, the LSTM-based models show the largest performance improvements in the north-east of Scotland and in south-east of England. The south-east of England remained difficult to model, however, in part due to the inability of the LSTMs configured in this study to learn groundwater processes, human abstractions and complex percolation properties from the hydro-meteorological variables typically employed for hydrological modelling.
Highlights
Rainfall–runoff models have evolved over many decades, reflecting a diversity of applications and purposes
The Long short-term memory (LSTM) and Entity Aware LSTM (EA LSTM) models produce accurate simulations across Great Britain when evaluated using a variety of metrics, with differing levels of performance improvement over the benchmark conceptual models (See Table 3)
We have demonstrated that the LSTM is an effective model architecture for extracting information from hydrometeorological data, providing a data-driven benchmark showing what is achievable given the information contained in available observation data from CAMELS-Great Britain (GB) (Nearing et al, 2021)
Summary
Rainfall–runoff models have evolved over many decades, reflecting a diversity of applications and purposes. These models range from physically based, spatially explicit models such as SHETRAN (Birkinshaw et al, 2010), CLASSIC (Crooks et al, 2014) and PARFLOW (Maxwell et al, 2009) to lumped conceptual models such as TOPMODEL (Beven and Kirkby, 1979) and VIC (Liang, 1994). Data-driven models range from simple regression models to large neural networks with thousands of parameters. These methods draw on empirical relationships between inputs and outputs to form a representation of how the hydrological system operates more generally (Beven, 2011).
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
