The Use of Recurrent Neural Networks (S-RNN, LSTM, GRU) For Flood Forecasting Based on Data Extracted from Classical Hydraulic Modeling DOI Creative Commons

Andrei Mihai Rugină

Modelling in Civil Environmental Engineering, Год журнала: 2023, Номер 18(3), С. 1 - 18

Опубликована: Сен. 1, 2023

Abstract Floods are natural disasters that have a significant impact on everyday human life, both through material losses and loss of life. In the context climate change, these events may be more frequent or dangerous. For real-time flood forecasting, fast methods for determining hydrographs along watercourses needed. Classic hydraulic modeling software provides satisfactory results, but in many cases calculation time can high. Another approach, different from classical is use neural networks forecasting hydrographs. Thus, present study aims to analyze three types recurrent networks, including SRNN, RNN-LSTM, RNN-GRU. each network type, flow level resulting were provided as input training data. Using deep learning environment, based previous calibration validation 2 historical modeled. The obtained extremely close those recorded, while running tens times smaller.

Язык: Английский

Intensity-Duration-Frequency Curves and Precipitation Frequency Analysis for the Ogun-Osun River Basin Using Annual Maxima Approach DOI
Habeeb Adedotun Alabi, Temitayo Abayomi Ewemoje, Benard Juma

и другие.

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0

Accounting for historical data uncertainty in flood frequency analysis: the Upper Rhine River DOI Creative Commons
M. Lang, Jérôme Le Coz, Benjamin Renard

и другие.

Journal of Hydrology, Год журнала: 2025, Номер unknown, С. 133480 - 133480

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0

Improving the Consistency of Hydrologic Event Identification DOI Creative Commons
Mohammad Masoud Mohammadpour Khoie, Danlu Guo, Conrad Wasko

и другие.

Environmental Modelling & Software, Год журнала: 2025, Номер unknown, С. 106521 - 106521

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0

Delineation of flood risk terrains and rainfall visualisation in the North Western part of Ghana DOI
Benjamin Wullobayi Dekongmen, Amos T. Kabo‐bah, Geophrey K. Anornu

и другие.

Modeling Earth Systems and Environment, Год журнала: 2024, Номер 10(3), С. 4567 - 4594

Опубликована: Май 16, 2024

Язык: Английский

Процитировано

2

Impact of catchment and climate attributes on flood generating processes and their effect on flood statistics DOI Creative Commons
Svenja Fischer, Markus Pahlow, Shailesh Kumar Singh

и другие.

Journal of Hydrology, Год журнала: 2024, Номер 646, С. 132361 - 132361

Опубликована: Ноя. 16, 2024

Язык: Английский

Процитировано

1

Shifting cold regions streamflow regimes in North America affect flood frequency analysis DOI
Donald H. Burn, Paul H. Whitfield

Hydrological Sciences Journal, Год журнала: 2024, Номер 70(1), С. 51 - 70

Опубликована: Окт. 28, 2024

Over threshold flood events in seventy years of data from 202 reference hydrometric stations Canada and the United States were separated into nival, mixed, pluvial types. While Mann-Kendall trend tests showed few significant trends magnitude, changes type fraction found over time, with annual mean temperature, precipitation. Nival decreased frequency seventy-year period 16% sites, while mixed increased (5%, 12%). These indicate a shift nival towards more dominated systems. Fewer analysis against four climate indices. Flood using combined distribution approach three types resulted larger magnitude design flow estimates (median increase 20 – 30 %) comparison results considering to be single population.

Язык: Английский

Процитировано

0

The Use of Recurrent Neural Networks (S-RNN, LSTM, GRU) For Flood Forecasting Based on Data Extracted from Classical Hydraulic Modeling DOI Creative Commons

Andrei Mihai Rugină

Modelling in Civil Environmental Engineering, Год журнала: 2023, Номер 18(3), С. 1 - 18

Опубликована: Сен. 1, 2023

Abstract Floods are natural disasters that have a significant impact on everyday human life, both through material losses and loss of life. In the context climate change, these events may be more frequent or dangerous. For real-time flood forecasting, fast methods for determining hydrographs along watercourses needed. Classic hydraulic modeling software provides satisfactory results, but in many cases calculation time can high. Another approach, different from classical is use neural networks forecasting hydrographs. Thus, present study aims to analyze three types recurrent networks, including SRNN, RNN-LSTM, RNN-GRU. each network type, flow level resulting were provided as input training data. Using deep learning environment, based previous calibration validation 2 historical modeled. The obtained extremely close those recorded, while running tens times smaller.

Язык: Английский

Процитировано

0