Towards better water level-flow relationships

被引：0

作者：

Oppel, Henning ^{[1
]}

Hartung, Alexander ^{[2
]}

Neumann, Juliane ^{[1
]}

Mewes, Benjamin ^{[1
]}

机构：

[1] Okeanos Smart Data Solut GmbH, Viktoriastr 29, D-44787 Bochum, Germany

[2] Emschergenossenschaft Lippeverband, Kronprinzenstr 24, D-45128 Essen, Germany

来源：

WASSERWIRTSCHAFT | 2023年 / 113卷 / 7-8期

关键词：

Compendex;

D O I：

10.1007/s35147-023-1879-2

中图分类号：

TV21 [水资源调查与水利规划];

学科分类号：

081501 ;

摘要：

Continuous or seasonal changes at measurement cross sections, as well as hysteresis effects, can significantly affect the acquisition of flow values using stage-discharge (W- Q) relationships. In the present study, we investigated how machine learning can be used to counteract these problems. The use of ML-based dynamic W-Q relationships as well as the use of additive deep learning models was tested at gauging stations at the Emscher and Lippe rivers.

引用

页码：64 / 67

页数：4

共 6 条

[1] Random forests [J].

Breiman, L .

MACHINE LEARNING, 2001, 45 (01) :5-32

[2] How to construct recursive digital filters for baseflow separation [J].

Eckhardt, K .

HYDROLOGICAL PROCESSES, 2005, 19 (02) :507-515

[3]

Hastie T, 2009, ELEMENTS STAT LEARNI

[4] Rainfall-runoff modelling using Long Short-Term Memory (LSTM) networks [J].

Kratzert, Frederik ;

Klotz, Daniel ;

Brenner, Claire ;

Schulz, Karsten ;

Herrnegger, Mathew .

HYDROLOGY AND EARTH SYSTEM SCIENCES, 2018, 22 (11) :6005-6022

[5]

Oppel H, 2021, WASSERWIRTSCHAFT, V111, P38

[6]

Pedregosa F, 2011, J MACH LEARN RES, V12, P2825

← 1 →