干旱区生态环境知识资源中心(Arid): River Flow Modeling in Semi-Arid and Humid Regions Using an Integrated Method Based on LARS-WG and LSTM Models

Arid

DOI	10.1007/s11269-023-03527-8
	River Flow Modeling in Semi-Arid and Humid Regions Using an Integrated Method Based on LARS-WG and LSTM Models
	Roushangar, Kiyoumars; Abdelzad, Sadegh
通讯作者	Roushangar, K
来源期刊	WATER RESOURCES MANAGEMENT
ISSN	0920-4741
EISSN	1573-1650
出版年	2023
卷号	37 期号:10 页码:3813-3831
英文摘要	River flow or runoff is an important water flux that can pose great threats to water security because of changes in its timing, magnitude, and seasonality. In this study, runoff variations in two different climate conditions, including the semi-arid and humid climates, were assessed using an integrated method based on the LARS-WG and Long Short-term Memory (LSTM) models. Runoff variations were simulated for the base period of 2001-2020 and it was forecasted for the future period of 2021-2040. Climatic parameters were evaluated using the LARS-WG statistical model for the selected stations under the RCPs scenarios. Then, outputs obtained under the RCPs scenarios, including minimum temperature, maximum temperature, precipitation, and sunshine hours, were considered as the input of the LSTM model. Results showed that the amount of mean runoff in spring for semi-arid areas and in spring and winter for humid areas will decrease in the next twenty years compared to the base period. For other seasons in both climates, the amount of runoff will increase. The maximum rainfall increase under the RCP8.5 scenario was 66.66 mm for the humid region and 30.06 mm for the semi-arid region compared to the base period. Also, the maximum temperature increase for the semi-arid region was 1.18 degrees C under the RCP8.5 scenario, and for the humid region was 1.07 degrees C under the RCP2.6 scenario. Results showed that the LSTM method can successfully model the future river flow values and river flow related to the previous day has a significant impact on the modeling process.
英文关键词	Deep learning method LARS-WG Runoff Semi-arid region Temperature
类型	Article
语种	英语
收录类别	SCI-E
WOS记录号	WOS:000994083100002
WOS关键词	ARTIFICIAL NEURAL-NETWORK ; DISCHARGE COEFFICIENT ; CLIMATE ; UNCERTAINTY ; SENSITIVITY ; EVAPORATION ; SYSTEM
WOS类目	Engineering, Civil ; Water Resources
WOS研究方向	Engineering ; Water Resources
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/399140
推荐引用方式 GB/T 7714	Roushangar, Kiyoumars,Abdelzad, Sadegh. River Flow Modeling in Semi-Arid and Humid Regions Using an Integrated Method Based on LARS-WG and LSTM Models[J],2023,37(10):3813-3831.
APA	Roushangar, Kiyoumars,&Abdelzad, Sadegh.(2023).River Flow Modeling in Semi-Arid and Humid Regions Using an Integrated Method Based on LARS-WG and LSTM Models.WATER RESOURCES MANAGEMENT,37(10),3813-3831.
MLA	Roushangar, Kiyoumars,et al."River Flow Modeling in Semi-Arid and Humid Regions Using an Integrated Method Based on LARS-WG and LSTM Models".WATER RESOURCES MANAGEMENT 37.10(2023):3813-3831.