High-frequency monitoring and machine learning models for assessing synoptic variability of suspended sediment flow of a small urban river
Authors
- Sergey R. Chalov Lomonosov Moscow State University https://orcid.org/0000-0002-6937-7020
- Vsevolod M. Moreido Institute of Water Problems
- Irina S. Denisova МГУ имени М. В. Ломоносова
- Ivan A. Solonikov Lomonosov Moscow State University
DOI:
https://doi.org/10.34753/HS.2023.5.1.59+ Keywords
+ Abstract
The article presents the results of high-frequency monitoring of sediment flow of the largest right tributary of the Moscow River within the city - the Setun River, carried out in 2020–2021 by automatic recorders of optical turbidity and water level. Several recordings were involved from November 16, 2019 to March 19, 2020 (recording frequency 10 minutes), from March 1, 2021 to April 2021 and from July 2021 to October 2021 (recording frequency 1 minute). Under conditions of high anthropogenic load, the average turbidity of the Setun River water reaches 100 mg/l, which is 3–4 times higher than background values. There is a correlation between turbidity and water flow rates, which increases with increasing time lag from 0 to 50 hours. Based on combining measurements of water turbidity with characteristics of water level and flow, precipitation, and air temperature, as well as derivatives from them, 5 machine learning models were configured to predict sediment runoff with a resolution of 30 minutes and 1 day. The best results were demonstrated by the recurrent neural network LSTM RNN for forecasting daily values of water turbidity: with a root-mean-square error of 10.8 NTU, the time and magnitude of local increases in the turbidity peak were correctly determined.
+ Author Biographies
Doctor of Geography, Associate Professor of the Department of Land Hydrology, Faculty of Geography, Lomonosov Moscow State University Ph.D., Head of Laboratory, Senior Researcher, Department of River Basin Hydrology, Laboratory of Hydroinformatics, Water Problems Institute of the Russian Academy of Sciences Postgraduate student, Department of Land Hydrology, Faculty of Geography, Lomonosov Moscow State University; Engineer, Surface Water Department, Surface Water Modeling Laboratory, Water Problems Institute of the Russian Academy of Sciences Master's student of the department. land hydrology, Faculty of Geography, Lomonosov Moscow State University
Sergey R. Chalov
Vsevolod M. Moreido
Irina S. Denisova
Ivan A. Solonikov
+ References
Караушев А.В. Сток наносов, его изучение и географическое распределение. Л.: Гидрометеоиздат, 1977. 240 с.
Прогнозирование уровней воды на реках Краснодарского края с использованием методов машинного обучения / Белякова П.А., Морейдо В.М., Цыпленков А.С., Амербаев А.Н., Гречишникова Д.А., Курочкина Л.С., Филиппов В.А., Макеев М.С. // Водные ресурсы. 2022. Т. 49. № 1. С. 13–26. https://doi.org/10.31857/S0321059622010047.
Реакция водного стока малой городской реки на экстремальные дождевые осадки на территории Москвы в 2020 и 2021 гг. / Чалов С.Р., Платонов В.С., Морейдо В.М., Самохин М.А., Ярынич Ю.И., Коршунова Н.Н., Болгов М.В., Касимов Н.С. // Метеорология и гидрология. 2023. Т. 2. С. 69-79. https://doi.org/10.52002/0130-2906-2023-2-69-79.
Чалов С.Р., Цыпленков А.С. Роль крупномасштабной турбулентности в изменении мутности речных вод // Вестник Московского университета. Серия 5. География. 2020. №. 3. С. 34–46.
Черных О.Н., Сабитов М.А., Алтунин В.И. Типизированные приемы экологического восстановления малых рек Москвы (на примере р. Сетунь) // Природообустройство. 2015. №. 3. С. 57–64.
Behrouz M.S., Yazdi M.N., Sample D.J. Using Random Forest, a machine learning approach to predict nitrogen, phosphorus, and sediment event mean concentrations in urban runoff // Journal of Environmental Management. 2022. V. 317. P. 115412. https://doi.org/10.1016/j.jenvman.2022.115412.
Bhattacharya B., Solomatine D.P. Machine learning in sedimentation modelling // Neural Networks. 2006. V. 19. № 2. P. 208-214. https://doi.org/10.1016/j.neunet.2006.01.007.
Box G.E.P., Jenkins G.M. Time Series Analysis: Forecasting and Control Holden-Day series in time series analysis and digital processing. / San Francisco, CA: Holden-Day, 1970, 575 p.
Carling P.A., Orr H.G. Morphology of riffle–pool sequences in the River Severn, England // Earth Surface Processes and Landforms: The Journal of the British Geomorphological Research Group. 2000. V. 25. № 4. P. 369-384. https://doi.org/10.1002/(SICI)1096-9837(200004)25:4<369::AID-ESP60>3.0.CO;2-M.
Chalov S.R., Platonov V.S., Moreido V.M., Samokhin M.A. Small Urban River Runoff Response to 2020 and 2021 Extreme Rainfalls on the Territory of Moscow // Russian meteorology and hydrology. 2023. V. 48. № 2. P. 138-146. https://doi.org/10.3103/S1068373923020061.
Chen T., Guestrin C. XGBoost: A Scalable Tree Boosting System // Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining : KDD ’16. – New York, NY, USA: ACM, 2016. P. 785-794. https://doi.org/10.1145/2939672.2939785.
Cohen S., Kettner A.J., Syvitski J.P.M. Global suspended sediment and water discharge dynamics between 1960 and 2010: Continental trends and intra-basin sensitivity // Global and planetary change. 2014. V. 115. P. 44-58. https://doi.org/10.1016/j.gloplacha.2014.01.011.
Hochreiter S., Schmidhuber J. Long Short-Term Memory // Neural Computation. 1997. V. 9. № 8. P. 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735.
Moreido V., Gartsman B., Solomatine D.P., Suchilina Z. How Well Can Machine Learning Models Perform without Hydrologists? Application of Rational Feature Selection to Improve Hydrological Forecasting // Water. 2021. V. 13. № 12. P. 1696. https://doi.org/10.3390/w13121696.
On the complexities of sediment load modeling using integrative machine learning: Application of the great river of Loíza in Puerto Rico / Zounemat-Kermani M., Mahdavi-Meymand A., Alizamir M., Adarsh S., Yaseen Z.M. // Journal of Hydrology. 2020. V. 585. P. 124759. https://doi.org/10.1016/j.jhydrol.2020.124759.
Press W.H., Teukolsky S.A. Savitzky-Golay Smoothing Filters // Computers in Physics. 1990. V. 4. № 6. P. 669-672. https://doi.org/10.1063/1.4822961.
Rainfall-runoff modelling using Long Short-Term Memory (LSTM) networks / Kratzert F., Klotz D., Brenner C., Schulz K., Herrnegger M. // Hydrology and Earth System Sciences. 2018. V. 22. № 11. P. 6005-6022. https://doi.org/10.5194/hess-22-6005-2018.
Rstudio Team RStudio: Integrated development for R. RStudio, Inc., Boston MA // RStudio. 2019.
Scikit-learn: Machine Learning in Python. / Pedregosa F. et al. // the Journal of machine Learning research. 2011. V. 12. P. 2825-2830.
Sharafati A., Haji Seyed Asadollah S.B., Motta D., Yaseen Z.M. Application of newly developed ensemble machine learning models for daily suspended sediment load prediction and related uncertainty analysis // Hydrological Sciences Journal. 2020. V. 65. № 12. P. 2022-2042. https://doi.org/10.1080/02626667.2020.1786571.
Sokolov D.I., Erina O.N., Tereshina M.A., Puklakov V. V. Impact Of Mozhaysk Dam On The Moscow River Sediment Transport // GEOGRAPHY, ENVIRONMENT, SUSTAINABILITY. 2020. V. 13. №. 4. P. 24-31. https://doi.org/10.24057/2071-9388-2019-150.
Solomatine D.P., Ostfeld A. Data-driven modelling: some past experiences and new approaches // Journal of Hydroinformatics. 2008. V. 10. № 1. P. 3. https://doi.org/10.2166/hydro.2008.015.
Taylor S.J., Letham B. Forecasting at Scale // The American Statistician. 2018. V. 72. № 1. P. 37-45. https://doi.org/10.1080/00031305.2017.1380080.
+ Read article online
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Hydrosphere. Hazard processes and phenomena
This work is licensed under a Creative Commons Attribution 4.0 International License.