Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths

Abstract Accurate high-resolution runoff predictions are essential for effective flood mitigation and water planning. In hydrology, conceptual models are preferred for their simplicity, despite their limited capacity for accurate predictions. Deep-learning applications have recently shown promise fo...

Full description

Saved in:
Bibliographic Details
Main Authors: Mohamed M. Fathi, Md Abdullah Al Mehedi, Virginia Smith, Anjali M. Fernandes, Michael T. Hren, Dennis O. Terry
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-96577-4
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849312101866143744
author Mohamed M. Fathi
Md Abdullah Al Mehedi
Virginia Smith
Anjali M. Fernandes
Michael T. Hren
Dennis O. Terry
author_facet Mohamed M. Fathi
Md Abdullah Al Mehedi
Virginia Smith
Anjali M. Fernandes
Michael T. Hren
Dennis O. Terry
author_sort Mohamed M. Fathi
collection DOAJ
description Abstract Accurate high-resolution runoff predictions are essential for effective flood mitigation and water planning. In hydrology, conceptual models are preferred for their simplicity, despite their limited capacity for accurate predictions. Deep-learning applications have recently shown promise for runoff predictions; however, they usually require longer input data sequences, especially for high-temporal resolution simulations, thus leading to increased model complexity. To address these challenges, this study evaluates the robustness of two novel approaches using Long Short-Term Memory (LSTM) models. The first model integrates the outputs of a simple conceptual model with LSTM capabilities, while the second model is a stand-alone model that combines coarse and fine temporal inputs to capture both long and short dependencies. To ensure accuracy and reliability, we utilized a century-long meteorological dataset generated from a sophisticated physics-based model, eliminating any influence of measurement errors. The training phase employed multiple sub-periods ranging from 7- to 50-year, with a separate 50-year subset for validation. Our findings highlight the consistent improvement of both LSTM models with increasing training dataset lengths, while conceptual models show no notable enhancement beyond 15 years of training data. Both LSTM models demonstrate superior performance in capturing the reference flow duration curve, offering a promising pathway for more computationally efficient models for runoff predictions.
format Article
id doaj-art-4c87e38915a042f9bab0ca53bc7f9563
institution Kabale University
issn 2045-2322
language English
publishDate 2025-05-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-4c87e38915a042f9bab0ca53bc7f95632025-08-20T03:53:12ZengNature PortfolioScientific Reports2045-23222025-05-0115111210.1038/s41598-025-96577-4Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengthsMohamed M. Fathi0Md Abdullah Al Mehedi1Virginia Smith2Anjali M. Fernandes3Michael T. Hren4Dennis O. Terry5Department of Civil Engineering, Faculty of Engineering, Fayoum UniversityDepartment of Civil and Environmental Engineering, Villanova UniversityDepartment of Civil and Environmental Engineering, Villanova UniversityDepartment of Earth and Environmental Sciences, Denison UniversityDepartment of Earth Sciences, University of ConnecticutDepartment of Earth and Environmental Science, Temple UniversityAbstract Accurate high-resolution runoff predictions are essential for effective flood mitigation and water planning. In hydrology, conceptual models are preferred for their simplicity, despite their limited capacity for accurate predictions. Deep-learning applications have recently shown promise for runoff predictions; however, they usually require longer input data sequences, especially for high-temporal resolution simulations, thus leading to increased model complexity. To address these challenges, this study evaluates the robustness of two novel approaches using Long Short-Term Memory (LSTM) models. The first model integrates the outputs of a simple conceptual model with LSTM capabilities, while the second model is a stand-alone model that combines coarse and fine temporal inputs to capture both long and short dependencies. To ensure accuracy and reliability, we utilized a century-long meteorological dataset generated from a sophisticated physics-based model, eliminating any influence of measurement errors. The training phase employed multiple sub-periods ranging from 7- to 50-year, with a separate 50-year subset for validation. Our findings highlight the consistent improvement of both LSTM models with increasing training dataset lengths, while conceptual models show no notable enhancement beyond 15 years of training data. Both LSTM models demonstrate superior performance in capturing the reference flow duration curve, offering a promising pathway for more computationally efficient models for runoff predictions.https://doi.org/10.1038/s41598-025-96577-4Length of calibration datasetGR4H and GR5H modelsOne-step ahead predictionEfficient LSTM modelsHybrid ML and conceptual modelsLonger training dataset
spellingShingle Mohamed M. Fathi
Md Abdullah Al Mehedi
Virginia Smith
Anjali M. Fernandes
Michael T. Hren
Dennis O. Terry
Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
Scientific Reports
Length of calibration dataset
GR4H and GR5H models
One-step ahead prediction
Efficient LSTM models
Hybrid ML and conceptual models
Longer training dataset
title Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
title_full Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
title_fullStr Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
title_full_unstemmed Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
title_short Evaluation of LSTM vs. conceptual models for hourly rainfall runoff simulations with varied training period lengths
title_sort evaluation of lstm vs conceptual models for hourly rainfall runoff simulations with varied training period lengths
topic Length of calibration dataset
GR4H and GR5H models
One-step ahead prediction
Efficient LSTM models
Hybrid ML and conceptual models
Longer training dataset
url https://doi.org/10.1038/s41598-025-96577-4
work_keys_str_mv AT mohamedmfathi evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths
AT mdabdullahalmehedi evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths
AT virginiasmith evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths
AT anjalimfernandes evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths
AT michaelthren evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths
AT dennisoterry evaluationoflstmvsconceptualmodelsforhourlyrainfallrunoffsimulationswithvariedtrainingperiodlengths