Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model

<p>Flood forecasting systems play a key role in mitigating socioeconomic damage caused by flood events. The majority of these systems rely on process-based hydrologic models (PBHMs), which are used to predict future runoff. Many operational flood forecasting systems additionally implement mode...

Full description

Saved in:
Bibliographic Details
Main Authors: S. Gegenleithner, M. Pirker, C. Dorfmann, R. Kern, J. Schneider
Format: Article
Language:English
Published: Copernicus Publications 2025-04-01
Series:Hydrology and Earth System Sciences
Online Access:https://hess.copernicus.org/articles/29/1939/2025/hess-29-1939-2025.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850182921843900416
author S. Gegenleithner
S. Gegenleithner
M. Pirker
C. Dorfmann
R. Kern
J. Schneider
author_facet S. Gegenleithner
S. Gegenleithner
M. Pirker
C. Dorfmann
R. Kern
J. Schneider
author_sort S. Gegenleithner
collection DOAJ
description <p>Flood forecasting systems play a key role in mitigating socioeconomic damage caused by flood events. The majority of these systems rely on process-based hydrologic models (PBHMs), which are used to predict future runoff. Many operational flood forecasting systems additionally implement models aimed at enhancing the predictions of the PBHM, either by updating the PBHM's state variables in real time or by enhancing its forecasts in a post-processing step. For the latter, autoregressive integrated moving average (ARIMA) models are frequently employed. Despite their high popularity in flood forecasting, studies have pointed out potential shortcomings of ARIMA-type models, such as a decline in forecast accuracy with increasing lead time. In this study, we investigate the potential of long short-term memory (LSTM) networks for enhancing the forecast accuracy of an underperforming PBHM and evaluate whether they are able to overcome some of the challenges presented by ARIMA models. To achieve this, we developed two hindcast–forecast LSTM models and compared their forecast accuracies to that of a more conventional ARIMA model. To ensure comparability, one LSTM was restricted to use the same data as ARIMA (eLSTM), namely observed and simulated discharge, while the other additionally incorporated meteorologic forcings (PBHM-HLSTM). Considering the PBHM's poor performance, we further evaluated if the PBHM-HLSTM was able to extract valuable information from the PBHM's results by analyzing the relative importance of each input feature. Contrary to ARIMA, the LSTM networks were able to mostly sustain a high forecast accuracy for longer lead times. Furthermore, the PBHM-HLSTM also achieved a high prediction accuracy for flood events, which was not the case for ARIMA or the eLSTM. Our results also revealed that the PBHM-HLSTM relied, to some degree, on the PBHM's results, despite its mostly poor performance. Our results suggest that LSTM models, especially when provided with meteorologic forcings, offer a promising alternative to frequently employed ARIMA models in operational flood forecasting systems.</p>
format Article
id doaj-art-2a2b66871eec476ca606094ce42a97d1
institution OA Journals
issn 1027-5606
1607-7938
language English
publishDate 2025-04-01
publisher Copernicus Publications
record_format Article
series Hydrology and Earth System Sciences
spelling doaj-art-2a2b66871eec476ca606094ce42a97d12025-08-20T02:17:29ZengCopernicus PublicationsHydrology and Earth System Sciences1027-56061607-79382025-04-01291939196210.5194/hess-29-1939-2025Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic modelS. Gegenleithner0S. Gegenleithner1M. Pirker2C. Dorfmann3R. Kern4J. Schneider5Institute of Hydraulic Engineering and Water Resources Management, Graz University of Technology, Stremayrgasse 10/II, 8010 Graz, Austriaflow engineering, Lessingstraße 30, 8010 Graz, AustriaInstitute of Hydraulic Engineering and Water Resources Management, Graz University of Technology, Stremayrgasse 10/II, 8010 Graz, Austriaflow engineering, Lessingstraße 30, 8010 Graz, AustriaInstitute of Interactive Systems and Data Science, Graz University of Technology, Sandgasse 36/III, 8010 Graz, AustriaInstitute of Hydraulic Engineering and Water Resources Management, Graz University of Technology, Stremayrgasse 10/II, 8010 Graz, Austria<p>Flood forecasting systems play a key role in mitigating socioeconomic damage caused by flood events. The majority of these systems rely on process-based hydrologic models (PBHMs), which are used to predict future runoff. Many operational flood forecasting systems additionally implement models aimed at enhancing the predictions of the PBHM, either by updating the PBHM's state variables in real time or by enhancing its forecasts in a post-processing step. For the latter, autoregressive integrated moving average (ARIMA) models are frequently employed. Despite their high popularity in flood forecasting, studies have pointed out potential shortcomings of ARIMA-type models, such as a decline in forecast accuracy with increasing lead time. In this study, we investigate the potential of long short-term memory (LSTM) networks for enhancing the forecast accuracy of an underperforming PBHM and evaluate whether they are able to overcome some of the challenges presented by ARIMA models. To achieve this, we developed two hindcast–forecast LSTM models and compared their forecast accuracies to that of a more conventional ARIMA model. To ensure comparability, one LSTM was restricted to use the same data as ARIMA (eLSTM), namely observed and simulated discharge, while the other additionally incorporated meteorologic forcings (PBHM-HLSTM). Considering the PBHM's poor performance, we further evaluated if the PBHM-HLSTM was able to extract valuable information from the PBHM's results by analyzing the relative importance of each input feature. Contrary to ARIMA, the LSTM networks were able to mostly sustain a high forecast accuracy for longer lead times. Furthermore, the PBHM-HLSTM also achieved a high prediction accuracy for flood events, which was not the case for ARIMA or the eLSTM. Our results also revealed that the PBHM-HLSTM relied, to some degree, on the PBHM's results, despite its mostly poor performance. Our results suggest that LSTM models, especially when provided with meteorologic forcings, offer a promising alternative to frequently employed ARIMA models in operational flood forecasting systems.</p>https://hess.copernicus.org/articles/29/1939/2025/hess-29-1939-2025.pdf
spellingShingle S. Gegenleithner
S. Gegenleithner
M. Pirker
C. Dorfmann
R. Kern
J. Schneider
Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
Hydrology and Earth System Sciences
title Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
title_full Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
title_fullStr Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
title_full_unstemmed Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
title_short Long short-term memory networks for enhancing real-time flood forecasts: a case study for an underperforming hydrologic model
title_sort long short term memory networks for enhancing real time flood forecasts a case study for an underperforming hydrologic model
url https://hess.copernicus.org/articles/29/1939/2025/hess-29-1939-2025.pdf
work_keys_str_mv AT sgegenleithner longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel
AT sgegenleithner longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel
AT mpirker longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel
AT cdorfmann longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel
AT rkern longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel
AT jschneider longshorttermmemorynetworksforenhancingrealtimefloodforecastsacasestudyforanunderperforminghydrologicmodel