Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions

This study examines streamflow simulations using deep learning (DL) to understand the information extraction capability of global DL models trained on multiple watersheds. The study separately examined the entire streamflow time series and recession flow predictions. It introduces a global–local (GL...

Full description

Saved in:
Bibliographic Details
Main Authors: Abhinav Gupta, Sean A. McKenna
Format: Article
Language:English
Published: Elsevier 2025-01-01
Series:Journal of Hydrology X
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2589915524000282
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850249539871571968
author Abhinav Gupta
Sean A. McKenna
author_facet Abhinav Gupta
Sean A. McKenna
author_sort Abhinav Gupta
collection DOAJ
description This study examines streamflow simulations using deep learning (DL) to understand the information extraction capability of global DL models trained on multiple watersheds. The study separately examined the entire streamflow time series and recession flow predictions. It introduces a global–local (GL) modeling strategy, where the global model outputs are fed as input to a locally trained model, with the hypothesis that the local model can leverage watershed-specific information that the global model may miss. The GL models demonstrate enhanced accuracy in recession flow prediction for 20-30% of the watersheds compared to the global and local models. However, considering the entire hydrograph, the GL models often perform worse than the global model. Further, the DL models were trained on two different objective functions. The performance of the global model in a watershed depended strongly upon the objective function used. These results suggest that the performance of global models is affected by watershed uniqueness, suggesting that even a global DL model should be tailored to individual watersheds for optimal performance.
format Article
id doaj-art-5ba24535108c4dcaa26072b7b98d9aa8
institution OA Journals
issn 2589-9155
language English
publishDate 2025-01-01
publisher Elsevier
record_format Article
series Journal of Hydrology X
spelling doaj-art-5ba24535108c4dcaa26072b7b98d9aa82025-08-20T01:58:28ZengElsevierJournal of Hydrology X2589-91552025-01-012610019810.1016/j.hydroa.2024.100198Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functionsAbhinav Gupta0Sean A. McKenna1Department of Chemical and Environmental Engineering, University of Cincinnati, Cincinnati, OH, USA; Corresponding author at: Department of Chemical and Environmental Engineering, 2600 Clifton Ave., University of Cincinnati, Cincinnati, OH 45221, USA.Division of Hydrologic Sciences, Desert Research Institute, Reno, NV, USAThis study examines streamflow simulations using deep learning (DL) to understand the information extraction capability of global DL models trained on multiple watersheds. The study separately examined the entire streamflow time series and recession flow predictions. It introduces a global–local (GL) modeling strategy, where the global model outputs are fed as input to a locally trained model, with the hypothesis that the local model can leverage watershed-specific information that the global model may miss. The GL models demonstrate enhanced accuracy in recession flow prediction for 20-30% of the watersheds compared to the global and local models. However, considering the entire hydrograph, the GL models often perform worse than the global model. Further, the DL models were trained on two different objective functions. The performance of the global model in a watershed depended strongly upon the objective function used. These results suggest that the performance of global models is affected by watershed uniqueness, suggesting that even a global DL model should be tailored to individual watersheds for optimal performance.http://www.sciencedirect.com/science/article/pii/S2589915524000282StreamflowRecession flowsHydrological modelingDeep learningWatershed uniquenessUncertainty
spellingShingle Abhinav Gupta
Sean A. McKenna
Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
Journal of Hydrology X
Streamflow
Recession flows
Hydrological modeling
Deep learning
Watershed uniqueness
Uncertainty
title Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
title_full Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
title_fullStr Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
title_full_unstemmed Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
title_short Hydrograph and recession flows simulations using deep learning: Watershed uniqueness and objective functions
title_sort hydrograph and recession flows simulations using deep learning watershed uniqueness and objective functions
topic Streamflow
Recession flows
Hydrological modeling
Deep learning
Watershed uniqueness
Uncertainty
url http://www.sciencedirect.com/science/article/pii/S2589915524000282
work_keys_str_mv AT abhinavgupta hydrographandrecessionflowssimulationsusingdeeplearningwatersheduniquenessandobjectivefunctions
AT seanamckenna hydrographandrecessionflowssimulationsusingdeeplearningwatersheduniquenessandobjectivefunctions